You searched for subject:(Bayesian Inference).
Showing records 1 – 30 of
823 total matches.
◁ [1] [2] [3] [4] [5] … [28] ▶
1.
Doraiswamy, Srikrishna.
Characterization of Nonlinear Material Response in the Presence of Large Uncertainties ??? A Bayesian Approach.
Degree: 2013, Texas Digital Library
URL: http://hdl.handle.net/1969; 
http://hdl.handle.net/2249.1/66804
► The aim of the current work is to develop a Bayesian approach to model and simulate the behavior of materials with nonlinear mechanical response in…
(more)
▼ The aim of the current work is to develop a
Bayesian approach to model and simulate the behavior of materials with nonlinear mechanical response in the presence of significant uncertainties in the experimental data as well as the applicability of models. The core idea of this approach is to combine deterministic approaches by the use of physics based models, with ideas from
Bayesian inference to account for such uncertainties.
Traditionally, parameters of models in mechanics have been identified through deterministic approaches to obtain single point estimates. Such methods perform very well for linear models and are the preferred approach in identifying model parameters, especially for precisely engineered systems such as structures and machinery. But in the presence of large variations such as in the response of biological materials, such deterministic approaches do not sufficiently capture the uncertainty in the response. We propose that the model parameters need to encode the spread that is observed in the data in addition to modeling the physics of the system. To this end, we propose the idea of probability distributions for model parameters in order to incorporate the uncertainty in the data.
We demonstrate this probabilistic approach to identifying model parameters with the example of two problems: the characterization of sheep arteries using data from inflation experiments and the problem of detecting an inhomogeneity in a cantilever beam. The parameters in the artery characterization problem are the model parameters in the constitutive models and in the cantilever problem the parameters are the stiffnesses of the inhomogeneity and the material of the beam. For each of these problems, we compute the probability distribution of the parameters using
Bayesian inference.
We show that the probability distributions of parameters can be used towards two kinds of diagnostics: assigning probability to a hypothesis (inhomogeneity detection problem) and using the probability distribution for classifying newly obtained data (characterization of artery data). For the inhomogeneity detection problem, the hypothesis is a statement on the ratio of the stiffnesses and it is observed that the probability of the hypothesis matches well with the data. In the case of the artery characterization problem, new data was successfully classified using the probability distributions computed with training data.
Advisors/Committee Members: Srinivasa, Arun R (advisor).
Subjects/Keywords: Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Doraiswamy, S. (2013). Characterization of Nonlinear Material Response in the Presence of Large Uncertainties ??? A Bayesian Approach. (Thesis). Texas Digital Library. Retrieved from http://hdl.handle.net/1969; http://hdl.handle.net/2249.1/66804
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Doraiswamy, Srikrishna. “Characterization of Nonlinear Material Response in the Presence of Large Uncertainties ??? A Bayesian Approach.” 2013. Thesis, Texas Digital Library. Accessed March 04, 2021.
http://hdl.handle.net/1969; http://hdl.handle.net/2249.1/66804.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Doraiswamy, Srikrishna. “Characterization of Nonlinear Material Response in the Presence of Large Uncertainties ??? A Bayesian Approach.” 2013. Web. 04 Mar 2021.
Vancouver:
Doraiswamy S. Characterization of Nonlinear Material Response in the Presence of Large Uncertainties ??? A Bayesian Approach. [Internet] [Thesis]. Texas Digital Library; 2013. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/1969; http://hdl.handle.net/2249.1/66804.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Doraiswamy S. Characterization of Nonlinear Material Response in the Presence of Large Uncertainties ??? A Bayesian Approach. [Thesis]. Texas Digital Library; 2013. Available from: http://hdl.handle.net/1969; http://hdl.handle.net/2249.1/66804
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
2.
Sugden, Lauren Alpert.
Structure, Variation, and Reproducibility: Bayesian
inference in problems arising from the study of RNA and an
RNA-binding protein.
Degree: PhD, Applied Mathematics, 2014, Brown University
URL: https://repository.library.brown.edu/studio/item/bdr:386269/
► Far from being solely a passive messenger between DNA and protein, RNA is a complex molecule involved in regulation at many levels. While the best-known…
(more)
▼ Far from being solely a passive messenger between DNA
and protein, RNA is a complex molecule involved in regulation at
many levels. While the best-known RNAs, messenger RNAs (mRNA), do
exist to carry information, there is also a large population of
non-coding RNAs that can take on complicated structures that enable
them to perform functions throughout the cell. Even mRNAs are not
static molecules, however. Mechanisms for altering mRNA sequences
result in transcripts that are no longer faithful representations
of the information in the genome. One such mechanism is RNA editing
by ADAR, which binds double-stranded RNA and targets a particular
adenosine for conversion to inosine, which is recognized by the
cell as guanosine. A major consequence of editing is amino-acid
recoding, resulting in protein diversification. In this thesis, we
look at three problems motivated by the study of RNA and ADAR.
First, we propose a method for assessing the reproducibility of
genome-scale studies that make a large number of predictions, using
the prediction of ADAR binding sites as a motivating example. We
then address the problem of avoiding false positives when
identifying subtle signals such as ADAR modifications in
high-throughput sequencing data in the presence of genetic
polymorphisms specific to laboratory populations. Finally, we turn
to structural prediction of RNA, inferring the common structural
and sequence characteristics of a set of related
transcripts.
Advisors/Committee Members: Lawrence, Charles (Director), Thompson, William (Reader), Reenan, Robert (Reader).
Subjects/Keywords: Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Sugden, L. A. (2014). Structure, Variation, and Reproducibility: Bayesian
inference in problems arising from the study of RNA and an
RNA-binding protein. (Doctoral Dissertation). Brown University. Retrieved from https://repository.library.brown.edu/studio/item/bdr:386269/
Chicago Manual of Style (16th Edition):
Sugden, Lauren Alpert. “Structure, Variation, and Reproducibility: Bayesian
inference in problems arising from the study of RNA and an
RNA-binding protein.” 2014. Doctoral Dissertation, Brown University. Accessed March 04, 2021.
https://repository.library.brown.edu/studio/item/bdr:386269/.
MLA Handbook (7th Edition):
Sugden, Lauren Alpert. “Structure, Variation, and Reproducibility: Bayesian
inference in problems arising from the study of RNA and an
RNA-binding protein.” 2014. Web. 04 Mar 2021.
Vancouver:
Sugden LA. Structure, Variation, and Reproducibility: Bayesian
inference in problems arising from the study of RNA and an
RNA-binding protein. [Internet] [Doctoral dissertation]. Brown University; 2014. [cited 2021 Mar 04].
Available from: https://repository.library.brown.edu/studio/item/bdr:386269/.
Council of Science Editors:
Sugden LA. Structure, Variation, and Reproducibility: Bayesian
inference in problems arising from the study of RNA and an
RNA-binding protein. [Doctoral Dissertation]. Brown University; 2014. Available from: https://repository.library.brown.edu/studio/item/bdr:386269/
3.
Lin, Luan.
Bayesian Inference and High-D space Characterization with
application in Paleoclimatology and Biology.
Degree: PhD, Applied Mathematics, 2012, Brown University
URL: https://repository.library.brown.edu/studio/item/bdr:297525/
► This thesis is a mathematical study of paleoclimatology and computational biology. Part I gives an introduction and overview of this dissertation. Part II presents the…
(more)
▼ This thesis is a mathematical study of
paleoclimatology and computational biology.
Part I gives an introduction and overview of this
dissertation.
Part II presents the applications of hidden Markov model in
paleoclimatology study.
Chapter 1 employs a two state hidden Markov model with
multivariate emission to challenge the assumption that ENSO-like
variability dominated Peru margin oceanography on decadal and
longer time scales over the Holocene epoch. The HMM result shows
that two regimes of variability dominate, one shows strong
correlations between surface and subsurface proxies while the other
does not.
Chapter 2 is another application of hidden Markov model in
paleoclimatology. A pair hidden Markov model is built to implement
the alignments between pairs of stratigraphic records and provide
uncertainty analysis. In addition to the most probable alignment,
centroid alignment is also investigated.
Part III focuses on characterization of high dimensional
space.
Chapter 3 develops an iterative algorithm to characterize the
high dimensional spaces by identifying the most informative
variables through mutual information based criteria. Application to
the posterior space of RNA secondary structure is presented.
Chapter 4 proposes a novel model based clustering algorithm
based on L
p distance in the situations where not only samples
but also corresponding densities/probabilities (or
subject to a
normalization constant) are available. We demonstrate the abilities
of this approach by experimenting via Gaussian Mixture Models.
Part IV makes some conclusions and suggests future
directions.
Advisors/Committee Members: Lawrence, Charles (Director), Herbert, Timothy (Reader), Thompson, William (Reader).
Subjects/Keywords: bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Lin, L. (2012). Bayesian Inference and High-D space Characterization with
application in Paleoclimatology and Biology. (Doctoral Dissertation). Brown University. Retrieved from https://repository.library.brown.edu/studio/item/bdr:297525/
Chicago Manual of Style (16th Edition):
Lin, Luan. “Bayesian Inference and High-D space Characterization with
application in Paleoclimatology and Biology.” 2012. Doctoral Dissertation, Brown University. Accessed March 04, 2021.
https://repository.library.brown.edu/studio/item/bdr:297525/.
MLA Handbook (7th Edition):
Lin, Luan. “Bayesian Inference and High-D space Characterization with
application in Paleoclimatology and Biology.” 2012. Web. 04 Mar 2021.
Vancouver:
Lin L. Bayesian Inference and High-D space Characterization with
application in Paleoclimatology and Biology. [Internet] [Doctoral dissertation]. Brown University; 2012. [cited 2021 Mar 04].
Available from: https://repository.library.brown.edu/studio/item/bdr:297525/.
Council of Science Editors:
Lin L. Bayesian Inference and High-D space Characterization with
application in Paleoclimatology and Biology. [Doctoral Dissertation]. Brown University; 2012. Available from: https://repository.library.brown.edu/studio/item/bdr:297525/
Louisiana State University
4.
Bhale, Ishan Singh.
Bayesian inference application to burglary detection.
Degree: MSCS, Computer Sciences, 2012, Louisiana State University
URL: etd-01222013-155000
;
https://digitalcommons.lsu.edu/gradschool_theses/2382
► Real time motion tracking is very important for video analytics. But very little research has been done in identifying the top-level plans behind the atomic…
(more)
▼ Real time motion tracking is very important for video analytics. But very little research has been done in identifying the top-level plans behind the atomic activities evident in various surveillance footages [61]. Surveillance videos can contain high level plans in the form of complex activities [61]. These complex activities are usually a combination of various articulated activities like breaking windshield, digging, and non-articulated activities like walking, running. We have developed a Bayesian framework for recognizing complex activities like burglary. This framework (belief network) is based on an expectation propagation algorithm [8] for approximate Bayesian inference. We provide experimental results showing the application of our framework for automatically detecting burglary from surveillance videos in real time.
Subjects/Keywords: Bayesian Inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Bhale, I. S. (2012). Bayesian inference application to burglary detection. (Masters Thesis). Louisiana State University. Retrieved from etd-01222013-155000 ; https://digitalcommons.lsu.edu/gradschool_theses/2382
Chicago Manual of Style (16th Edition):
Bhale, Ishan Singh. “Bayesian inference application to burglary detection.” 2012. Masters Thesis, Louisiana State University. Accessed March 04, 2021.
etd-01222013-155000 ; https://digitalcommons.lsu.edu/gradschool_theses/2382.
MLA Handbook (7th Edition):
Bhale, Ishan Singh. “Bayesian inference application to burglary detection.” 2012. Web. 04 Mar 2021.
Vancouver:
Bhale IS. Bayesian inference application to burglary detection. [Internet] [Masters thesis]. Louisiana State University; 2012. [cited 2021 Mar 04].
Available from: etd-01222013-155000 ; https://digitalcommons.lsu.edu/gradschool_theses/2382.
Council of Science Editors:
Bhale IS. Bayesian inference application to burglary detection. [Masters Thesis]. Louisiana State University; 2012. Available from: etd-01222013-155000 ; https://digitalcommons.lsu.edu/gradschool_theses/2382
University of Adelaide
5.
Gray, Caitlin.
On the application of Bayesian inference to network estimation problems.
Degree: 2020, University of Adelaide
URL: http://hdl.handle.net/2440/129332
► Interconnected network structures play a crucial role in many aspects of our lives. Understanding these networks and the dynamic processes that can propagate over them…
(more)
▼ Interconnected network structures play a crucial role in many aspects of our lives. Understanding these networks and the dynamic processes that can propagate over them gives rise to many interesting questions. While many network and dynamical models are probabilistic, many of the current methods to solve key network science problems do not account for the uncertainties in the systems. In this thesis we look at networks in a probabilistic setting and use
Bayesian methods to simulate from distributions of these networks to address key network science problems. This allows deeper understanding about the system of interest and allows us to quantify uncertainty for improved decision making. Specifically, in Chapters 3 and 4 we develop an algorithm that simulates connected random networks and explore how variations to the algorithm can improve performance of sampling the high dimensional posterior. This develops techniques that we further apply in Chapters 5 and 6 to an interesting inverse problem in social network analysis - the ‘Network
Inference problem’. The underlying premise is to infer network structure, how people are connected, when we can only observe things moving between actors in the social network, e.g., tweets of news articles. By using a
Bayesian method we can provide uncertainty estimates for the inferences we make. We consider a variety of methods to extend the applicability to a wide range of data types, including streaming data, and test the
inference on both simulated and real data. Finally, in Chapter 7, we consider the ‘Network Tomography problem’, which aims to infer node properties when we only have information about paths in the network. We highlight the benefits of
Bayesian inference methods in two specific applications to aid decision making when identifying routing policies over the internet. Throughout this work we highlight that applying
Bayesian inference techniques to novel applications in network science can expand the types of networks we can simulate, provide uncertainty quantification for making informed decisions and gain results and insight in low and messy data regimes.
Advisors/Committee Members: Roughan, Matthew (advisor), Mitchell, Lewis (advisor), School of Mathematical Sciences (school).
Subjects/Keywords: Network inference; information cascades; Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Gray, C. (2020). On the application of Bayesian inference to network estimation problems. (Thesis). University of Adelaide. Retrieved from http://hdl.handle.net/2440/129332
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Gray, Caitlin. “On the application of Bayesian inference to network estimation problems.” 2020. Thesis, University of Adelaide. Accessed March 04, 2021.
http://hdl.handle.net/2440/129332.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Gray, Caitlin. “On the application of Bayesian inference to network estimation problems.” 2020. Web. 04 Mar 2021.
Vancouver:
Gray C. On the application of Bayesian inference to network estimation problems. [Internet] [Thesis]. University of Adelaide; 2020. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2440/129332.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Gray C. On the application of Bayesian inference to network estimation problems. [Thesis]. University of Adelaide; 2020. Available from: http://hdl.handle.net/2440/129332
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Texas A&M University
6.
Das, Roneet.
Probabilistic Slope Stability Assessment of Submarine and Slides by the Use of Bayesian Inference.
Degree: MS, Civil Engineering, 2016, Texas A&M University
URL: http://hdl.handle.net/1969.1/157070
► Estimates of probability of slope failure based on Monte-Carlo methods depend upon the state of evidence on the slope stability model parameters. The Bayesian framework…
(more)
▼ Estimates of probability of slope failure based on Monte-Carlo methods depend upon the state of evidence on the slope stability model parameters. The
Bayesian framework illustrated in this paper to estimate the probability of failure against submarine landslide incorporates experimental data (undrained shear strength) with the initial state of evidence on the model parameters to achieve more certain and accurate model predictions and estimates of probability of failure.
The objective of this research was to determine the probability of failure of a submarine slope due to static loading conditions for a given state of evidence (e.g. soil data, slope stability model and expert beliefs). A physics-based forward model (infinite slope) was adopted to evaluate the probability of failure against sliding. The geotechnical and geometric parameters (unit weight of the slice, thickness, pseudo-static seismic coefficient and slope angle) of the proposed model were regarded as random variables. The initial state of evidence and level of uncertainty associated with the proposed model parameters were presented as prior probability distribution functions (log-normal distribution). The
Bayesian framework was adopted to calibrate the proposed model with synthetically generated experimental observations representing different in-situ undrained shear strength conditions.
Model predictions on the mobilized shear strength when sampled from posterior distributions of the model parameters showed greater certainty and accuracy with respect to the Monte-Carlo forward model simulations based on the prior distributions. Results showed significant changes in the landslide probability with the increase in amount of data for two scenarios used for model calibration, while indicating the correlation structure changes among the model parameters. This allowed to estimating the sampling scenarios and their corresponding confidence gains prior to a field investigation.
Advisors/Committee Members: Medina-Cetina, Zenon (advisor), Aubeny, Charles (committee member), Kim, Jihoon (committee member).
Subjects/Keywords: Slope Stability; Bayesian Inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Das, R. (2016). Probabilistic Slope Stability Assessment of Submarine and Slides by the Use of Bayesian Inference. (Masters Thesis). Texas A&M University. Retrieved from http://hdl.handle.net/1969.1/157070
Chicago Manual of Style (16th Edition):
Das, Roneet. “Probabilistic Slope Stability Assessment of Submarine and Slides by the Use of Bayesian Inference.” 2016. Masters Thesis, Texas A&M University. Accessed March 04, 2021.
http://hdl.handle.net/1969.1/157070.
MLA Handbook (7th Edition):
Das, Roneet. “Probabilistic Slope Stability Assessment of Submarine and Slides by the Use of Bayesian Inference.” 2016. Web. 04 Mar 2021.
Vancouver:
Das R. Probabilistic Slope Stability Assessment of Submarine and Slides by the Use of Bayesian Inference. [Internet] [Masters thesis]. Texas A&M University; 2016. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/1969.1/157070.
Council of Science Editors:
Das R. Probabilistic Slope Stability Assessment of Submarine and Slides by the Use of Bayesian Inference. [Masters Thesis]. Texas A&M University; 2016. Available from: http://hdl.handle.net/1969.1/157070
University of Otago
7.
Holmes, Tom.
Exploring GPS Signal Data and Position Estimation Through Bayesian Inference
.
Degree: 2011, University of Otago
URL: http://hdl.handle.net/10523/2032
► A new approach to position determination using the Global Positioning System (GPS) has been developed where post processing of ultra-short sequences of captured GPS satellite…
(more)
▼ A new approach to position determination using the Global Positioning System (GPS) has been developed where post processing of ultra-short sequences of captured GPS satellite signal data can produce an estimate of receiver location. The new fixing scheme, labelled `FastFix' works by downloading GPS satellite information (ephemeris) data from the internet rather than decoding the same data (at a rate of 50bps) from each satellite's broadcast signal. Fewer than five milliseconds of signal data, sampled at around 8MHz, is processed in a similar manner to a traditional GPS receiver chip (C/A code phases and carrier Doppler shifts are found through standard signal processing techniques) before a least squares optimization is used to estimate receiver position and GPS time. This thesis builds on the FastFix scheme by analyzing the binary signal data (output from the GPS receiver chip) using
Bayesian inference with Markov chain Monte Carlo sampling. The approach presented in this thesis borrows ideas from the FastFix scheme to begin the position determination process and results in position estimates typically within 250 meters of true receiver location. The
Bayesian approach allows not only position and time estimates but also their relative uncertainty to be determined.
Advisors/Committee Members: Molteno, Timothy (advisor).
Subjects/Keywords: GPS;
Bayesian;
Animal Tracking;
Inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Holmes, T. (2011). Exploring GPS Signal Data and Position Estimation Through Bayesian Inference
. (Masters Thesis). University of Otago. Retrieved from http://hdl.handle.net/10523/2032
Chicago Manual of Style (16th Edition):
Holmes, Tom. “Exploring GPS Signal Data and Position Estimation Through Bayesian Inference
.” 2011. Masters Thesis, University of Otago. Accessed March 04, 2021.
http://hdl.handle.net/10523/2032.
MLA Handbook (7th Edition):
Holmes, Tom. “Exploring GPS Signal Data and Position Estimation Through Bayesian Inference
.” 2011. Web. 04 Mar 2021.
Vancouver:
Holmes T. Exploring GPS Signal Data and Position Estimation Through Bayesian Inference
. [Internet] [Masters thesis]. University of Otago; 2011. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/10523/2032.
Council of Science Editors:
Holmes T. Exploring GPS Signal Data and Position Estimation Through Bayesian Inference
. [Masters Thesis]. University of Otago; 2011. Available from: http://hdl.handle.net/10523/2032
University of Miami
8.
Abeyruwan, Saminda.
PrOntoLearn: Unsupervised Lexico-Semantic Ontology Generation using Probabilistic Methods.
Degree: MS, Computer Science (Arts and Sciences), 2010, University of Miami
URL: https://scholarlyrepository.miami.edu/oa_theses/28
► An ontology is a formal, explicit specification of a shared conceptualization. Formalizing an ontology for a domain is a tedious and cumbersome process. It…
(more)
▼ An ontology is a formal, explicit specification of a shared conceptualization. Formalizing an ontology for a domain is a tedious and cumbersome process. It is constrained by the knowledge acquisition bottleneck (KAB). There exists a large number of text corpora that can be used for classification in order to create ontologies with the intention to provide better support for the intended parties. In our research we provide a novel unsupervised bottom-up ontology generation method. This method is based on lexico-semantic structures and
Bayesian reasoning to expedite the ontology generation process. This process also provides evidence to domain experts to build ontologies based on top-down approaches.
Advisors/Committee Members: Ubbo Visser, Geoff Sutcliffe, Stephan Schuerer.
Subjects/Keywords: An Ontology; Learning; Bayesian Inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Abeyruwan, S. (2010). PrOntoLearn: Unsupervised Lexico-Semantic Ontology Generation using Probabilistic Methods. (Thesis). University of Miami. Retrieved from https://scholarlyrepository.miami.edu/oa_theses/28
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Abeyruwan, Saminda. “PrOntoLearn: Unsupervised Lexico-Semantic Ontology Generation using Probabilistic Methods.” 2010. Thesis, University of Miami. Accessed March 04, 2021.
https://scholarlyrepository.miami.edu/oa_theses/28.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Abeyruwan, Saminda. “PrOntoLearn: Unsupervised Lexico-Semantic Ontology Generation using Probabilistic Methods.” 2010. Web. 04 Mar 2021.
Vancouver:
Abeyruwan S. PrOntoLearn: Unsupervised Lexico-Semantic Ontology Generation using Probabilistic Methods. [Internet] [Thesis]. University of Miami; 2010. [cited 2021 Mar 04].
Available from: https://scholarlyrepository.miami.edu/oa_theses/28.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Abeyruwan S. PrOntoLearn: Unsupervised Lexico-Semantic Ontology Generation using Probabilistic Methods. [Thesis]. University of Miami; 2010. Available from: https://scholarlyrepository.miami.edu/oa_theses/28
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
University of Cambridge
9.
O'Keeffe, Jonathan.
Mathematical models of the representation of faces in humans.
Degree: PhD, 2020, University of Cambridge
URL: https://doi.org/10.17863/CAM.46702
;
https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.793085
► The representation of faces is a crucial function of the human CNS, as demonstrated by the severe social difficulties experienced when people lack this ability…
(more)
▼ The representation of faces is a crucial function of the human CNS, as demonstrated by the severe social difficulties experienced when people lack this ability (prosopagnosia). However, the precise way in which faces are represented and differentiated from one another is not well understood. This work addresses two substantial issues. Firstly, how is information about faces integrated over time? In chapter 2 a simple model of temporal integration is set forth, based on the statistical technique of exponential smoothing. In chapter 3 results of experiments testing this model are presented, demonstrating the model to be inadequate in certain respects. In particular a systematic bias towards the origin of face space is observed, a phenomenon referred to as "bowing". Chapter 4 contains a further model, which aims to show how this bowing could arise from a Bayesian inferential process. The second issue, addressed in chapter 5 of this thesis, is how well human judgements of facial similarity correspond to predictions made using Basel Face Space (BFS), a popular and widely used representation of faces from the field of computer vision. The degree of agreement is quantified using a novel experimental approach, and subsequently salient differences between the biological face space and BFS, including some original findings relating to isotropy or directionality, are demonstrated.
Subjects/Keywords: Faces; computer vision; Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
O'Keeffe, J. (2020). Mathematical models of the representation of faces in humans. (Doctoral Dissertation). University of Cambridge. Retrieved from https://doi.org/10.17863/CAM.46702 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.793085
Chicago Manual of Style (16th Edition):
O'Keeffe, Jonathan. “Mathematical models of the representation of faces in humans.” 2020. Doctoral Dissertation, University of Cambridge. Accessed March 04, 2021.
https://doi.org/10.17863/CAM.46702 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.793085.
MLA Handbook (7th Edition):
O'Keeffe, Jonathan. “Mathematical models of the representation of faces in humans.” 2020. Web. 04 Mar 2021.
Vancouver:
O'Keeffe J. Mathematical models of the representation of faces in humans. [Internet] [Doctoral dissertation]. University of Cambridge; 2020. [cited 2021 Mar 04].
Available from: https://doi.org/10.17863/CAM.46702 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.793085.
Council of Science Editors:
O'Keeffe J. Mathematical models of the representation of faces in humans. [Doctoral Dissertation]. University of Cambridge; 2020. Available from: https://doi.org/10.17863/CAM.46702 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.793085
University of Toronto
10.
Cheng, Vincent.
Modeling the Climatology of Tornado Occurrence using Bayesian Inference.
Degree: PhD, 2014, University of Toronto
URL: http://hdl.handle.net/1807/68452
► Our mechanistic understanding of tornadic environments has significantly improved by the recent technological enhancements in the detection of tornadoes as well as the advances of…
(more)
▼ Our mechanistic understanding of tornadic environments has significantly improved by the recent technological enhancements in the detection of tornadoes as well as the advances of numerical weather predictive modeling. Nonetheless, despite the decades of active research, prediction of tornado occurrence remains one of the most difficult problems in meteorological and climate science. In our efforts to develop predictive tools for tornado occurrence, there are a number of issues to overcome, such as the treatment of inconsistent tornado records, the consideration of suitable combination of atmospheric predictors, and the selection of appropriate resolution to accommodate the variability in time and space. In this dissertation, I address each of these topics by undertaking three empirical (statistical) modeling studies, where I examine the signature of different atmospheric factors influencing the tornado occurrence, the sampling biases in tornado observations, and the optimal spatiotemporal resolution for studying tornado occurrence. In the first study, I develop a novel
Bayesian statistical framework to assess the probability of tornado occurrence in Canada, in which the sampling bias of tornado observations and the linkage between lightning climatology and tornadogenesis are considered. The results produced reasonable probability estimates of tornado occurrence for the under-sampled areas in the model domain. The same study also delineated the geographical variability in the lightning-tornado relationship across Canada. In the second study, I present a novel modeling framework to examine the relative importance of several key atmospheric variables (e.g., convective available potential energy, 0-3 km storm-relative helicity, 0-6 km bulk wind difference, 0-tropopause vertical wind shear) on tornado activity in North America. I found that the variable quantifying the updraft strength is more important during the warm season, whereas the effects of wind-related variables are more uniform across seasons. The residual variability of the same modeling framework (a reflection of the fidelity of the statistical formulation considered) is subsequently used to delineate distinct geographical patterns of tornado activity. This piece of information provides the foundation for the
Bayesian hierarchical prognostic model presented in the third chapter of my dissertation. The results of the latter approach reinforce my earlier finding that the spatial variability of the annual and warm seasonal tornado occurrence is well explained by convective available potential energy and storm relative helicity alone, while vertical wind shear is better at reproducing the cool season tornado activity. The
Bayesian hierarchical modeling framework offers a promising methodological tool for understanding regional tornado environments and obtaining reliable predictions in North America.
Advisors/Committee Members: George, Arhonditsis, Geography.
Subjects/Keywords: Bayesian Inference; Tornadoes; 0368
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Cheng, V. (2014). Modeling the Climatology of Tornado Occurrence using Bayesian Inference. (Doctoral Dissertation). University of Toronto. Retrieved from http://hdl.handle.net/1807/68452
Chicago Manual of Style (16th Edition):
Cheng, Vincent. “Modeling the Climatology of Tornado Occurrence using Bayesian Inference.” 2014. Doctoral Dissertation, University of Toronto. Accessed March 04, 2021.
http://hdl.handle.net/1807/68452.
MLA Handbook (7th Edition):
Cheng, Vincent. “Modeling the Climatology of Tornado Occurrence using Bayesian Inference.” 2014. Web. 04 Mar 2021.
Vancouver:
Cheng V. Modeling the Climatology of Tornado Occurrence using Bayesian Inference. [Internet] [Doctoral dissertation]. University of Toronto; 2014. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/1807/68452.
Council of Science Editors:
Cheng V. Modeling the Climatology of Tornado Occurrence using Bayesian Inference. [Doctoral Dissertation]. University of Toronto; 2014. Available from: http://hdl.handle.net/1807/68452
University of Melbourne
11.
Li, Yuan.
Probabilistic models for aggregating crowdsourced annotations.
Degree: 2019, University of Melbourne
URL: http://hdl.handle.net/11343/227106
► This thesis explores aggregation methods for crowdsourced annotations. Crowdsourcing is a popular means of creating training and evaluation datasets for machine learning, e.g. used for…
(more)
▼ This thesis explores aggregation methods for crowdsourced annotations. Crowdsourcing is a popular means of creating training and evaluation datasets for machine learning, e.g. used for computer vision, natural language processing, and information retrieval, at a low cost and in a timely manner. However, due to low-quality annotations, individual workers cannot be wholly trusted to provide reliable annotations and consequently items are typically redundantly labelled by different workers, with labels aggregated subsequently. Although many aggregation methods have been proposed to jointly learn non-uniform weights to workers and infer the truth, the simplest aggregation method, majority voting (MV) which grants workers equal votes towards consensus, still predominates in practice.
To find explanations to the predominance of MV, we conduct extensive experiments of evaluation on 19 datasets and identify two shortcomings that prevent existing methods from being applied in practice over the simple MV. A key finding is that most methods don’t significantly outperform MV across all datasets. These methods may achieve higher mean accuracy than MV does but are also outperformed by MV on several datasets. A secondary shortcoming is that several methods require slow and cumbersome inference, which doesn’t scale to large datasets that are common in practice.
To address the identified shortcomings, we propose two novel aggregation methods both of which significantly outperform MV. The first is a Bayesian version of a weighted average model. It learns unknown voting weights of workers in a principled way by estimating their posterior, unlike existing weighted average models that rely on heuristic update rules or optimising handcrafted objectives. The second approach, complementary to the above, is another Bayesian model that captures the correlations between worker labels which most existing models completely ignore or assume don’t exist. Learning the correlations also helps the method achieve the highest mean accuracy among all methods compared in our experiments.
When applying aggregation methods in practice, it’s typically assumed that the only information we have is worker labels, but in many situations more information is available. For the setting where item content is available, e.g. feature vectors, we propose a novel model for aggregating binary labels using a Boltzmann machine prior to bias similar instances towards sharing the same label. We also show further gains by integrating a proposed active learning heuristic. We also consider a second, related, setting where instances are sentences, the task is annotating which words in the sentence denote a named entity, structural outputs from a few classifiers are given, and the goal is ensembling those classifiers. We discuss the strategy of adapting aggregation methods for crowdsourcing into this setting. We also discuss the effect of a few gold labels on truth inference and approaches for effectively leveraging gold labels.
Subjects/Keywords: crowdsourcing; probabilistic models; Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Li, Y. (2019). Probabilistic models for aggregating crowdsourced annotations. (Doctoral Dissertation). University of Melbourne. Retrieved from http://hdl.handle.net/11343/227106
Chicago Manual of Style (16th Edition):
Li, Yuan. “Probabilistic models for aggregating crowdsourced annotations.” 2019. Doctoral Dissertation, University of Melbourne. Accessed March 04, 2021.
http://hdl.handle.net/11343/227106.
MLA Handbook (7th Edition):
Li, Yuan. “Probabilistic models for aggregating crowdsourced annotations.” 2019. Web. 04 Mar 2021.
Vancouver:
Li Y. Probabilistic models for aggregating crowdsourced annotations. [Internet] [Doctoral dissertation]. University of Melbourne; 2019. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/11343/227106.
Council of Science Editors:
Li Y. Probabilistic models for aggregating crowdsourced annotations. [Doctoral Dissertation]. University of Melbourne; 2019. Available from: http://hdl.handle.net/11343/227106
University of Sydney
12.
Abeywardana, Sachinthaka.
Variational Inference in Generalised Hyperbolic and von Mises-Fisher Distributions
.
Degree: 2015, University of Sydney
URL: http://hdl.handle.net/2123/16504
► Most real world data are skewed, contain more than the set of real numbers, and have higher probabilities of extreme events occurring compared to a…
(more)
▼ Most real world data are skewed, contain more than the set of real numbers, and have higher probabilities of extreme events occurring compared to a normal distribution. In this thesis we explore two non-Gaussian distributions, the Generalised Hyperbolic Distribution (GHD) and, the von-Mises Fisher (vMF) Distribution. These distributions are studied in the context of 1) Regression in heavy tailed data, 2) Quantifying variance of functions with reference to finding relevant quantiles and, 3) Clustering data that lie on the surface of the sphere. Firstly, we extend Gaussian Processes (GPs) and use the Genralised Hyperbolic Processes as a prior on functions instead. This prior is more flexible than GPs and is especially able to model data that has high kurtosis. The method is based on placing a Generalised Inverse Gaussian prior over the signal variance, which yields a scalar mixture of GPs. We show how to perform inference efficiently for the predictive mean and variance, and use a variational EM method for learning. Secondly, the skewed extension of the GHD is studied with respect to quantile regression. An underlying GP prior on the quantile function is used to make the inference non-parametric, while the skewed GHD is used as the data likelihood. The skewed GHD has a single parameter alpha which states the required quantile. Similar variational methods as the first contribution is used to perform inference. Finally, vMF distributions are introduced in order to cluster spherical data. In the two previous contributions continuous scalar mixtures of Gaussians were used to make the inference process simpler. However, for clustering, a discrete number of vMF distributions are typically used. We propose a Dirichlet Process (DP) to infer the number of clusters in the spherical data setup. The framework is extended to incorporate a nested and a temporal clustering architecture. Throughout this thesis in many cases the posterior cannot be calculated in closed form. Variational Bayesian approximations are derived in this situation for efficient inference. In certain cases further lower bounding of the optimisation function is required in order to perform Variational Bayes. These bounds themselves are novel.
Subjects/Keywords: machine learning;
Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Abeywardana, S. (2015). Variational Inference in Generalised Hyperbolic and von Mises-Fisher Distributions
. (Thesis). University of Sydney. Retrieved from http://hdl.handle.net/2123/16504
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Abeywardana, Sachinthaka. “Variational Inference in Generalised Hyperbolic and von Mises-Fisher Distributions
.” 2015. Thesis, University of Sydney. Accessed March 04, 2021.
http://hdl.handle.net/2123/16504.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Abeywardana, Sachinthaka. “Variational Inference in Generalised Hyperbolic and von Mises-Fisher Distributions
.” 2015. Web. 04 Mar 2021.
Vancouver:
Abeywardana S. Variational Inference in Generalised Hyperbolic and von Mises-Fisher Distributions
. [Internet] [Thesis]. University of Sydney; 2015. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2123/16504.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Abeywardana S. Variational Inference in Generalised Hyperbolic and von Mises-Fisher Distributions
. [Thesis]. University of Sydney; 2015. Available from: http://hdl.handle.net/2123/16504
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
University of Sydney
13.
McCalman, Lachlan Robert.
Function Embeddings for Multi-modal Bayesian Inference
.
Degree: 2013, University of Sydney
URL: http://hdl.handle.net/2123/12031
► Tractable Bayesian inference is a fundamental challenge in robotics and machine learning. Standard approaches such as Gaussian process regression and Kalman filtering make strong Gaussianity…
(more)
▼ Tractable Bayesian inference is a fundamental challenge in robotics and machine learning. Standard approaches such as Gaussian process regression and Kalman filtering make strong Gaussianity assumptions about the underlying distributions. Such assumptions, however, can quickly break down when dealing with complex systems such as the dynamics of a robot or multi-variate spatial models. In this thesis we aim to solve Bayesian regression and filtering problems without making assumptions about the underlying distributions. We develop techniques to produce rich posterior representations for complex, multi-modal phenomena. Our work extends kernel Bayes' rule (KBR), which uses empirical estimates of distributions derived from a set of training samples and embeds them into a high-dimensional reproducing kernel Hilbert space (RKHS). Bayes' rule itself occurs on elements of this space. Our first contribution is the development of an efficient method for estimating posterior density functions from kernel Bayes' rule, applied to both filtering and regression. By embedding fixed-mean mixtures of component distributions, we can efficiently find an approximate pre-image by optimising the mixture weights using a convex quadratic program. The result is a complex, multi-modal posterior representation. Our next contributions are methods for estimating cumulative distributions and quantile estimates from the posterior embedding of kernel Bayes' rule. We examine a number of novel methods, including those based on our density estimation techniques, as well as directly estimating the cumulative through use of the reproducing property of RKHSs. Finally, we develop a novel method for scaling kernel Bayes' rule inference to large datasets, using a reduced-set construction optimised using the posterior likelihood. This method retains the ability to perform multi-output inference, as well as our earlier contributions to represent explicitly non-Gaussian posteriors and quantile estimates.
Subjects/Keywords: Statistics;
Inference;
Bayesian;
Machine Learning
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
McCalman, L. R. (2013). Function Embeddings for Multi-modal Bayesian Inference
. (Thesis). University of Sydney. Retrieved from http://hdl.handle.net/2123/12031
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
McCalman, Lachlan Robert. “Function Embeddings for Multi-modal Bayesian Inference
.” 2013. Thesis, University of Sydney. Accessed March 04, 2021.
http://hdl.handle.net/2123/12031.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
McCalman, Lachlan Robert. “Function Embeddings for Multi-modal Bayesian Inference
.” 2013. Web. 04 Mar 2021.
Vancouver:
McCalman LR. Function Embeddings for Multi-modal Bayesian Inference
. [Internet] [Thesis]. University of Sydney; 2013. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2123/12031.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
McCalman LR. Function Embeddings for Multi-modal Bayesian Inference
. [Thesis]. University of Sydney; 2013. Available from: http://hdl.handle.net/2123/12031
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
University of Alberta
14.
Gonzalez, Ruben.
Bayesian Methods for On-Line Gross Error Detection and
Compensation.
Degree: MS, Department of Chemical and Materials
Engineering, 2010, University of Alberta
URL: https://era.library.ualberta.ca/files/w6634487q
► Data reconciliation and gross error detection are traditional methods toward detecting mass balance inconsistency within process instrument data. These methods use a static approach for…
(more)
▼ Data reconciliation and gross error detection are
traditional methods toward detecting mass balance inconsistency
within process instrument data. These methods use a static approach
for statistical evaluation. This thesis is concerned with using an
alternative statistical approach (Bayesian statistics) to detect
mass balance inconsistency in real time. The proposed dynamic
Baysian solution makes use of a state space process model which
incorporates mass balance relationships so that a governing set of
mass balance variables can be estimated using a Kalman filter. Due
to the incorporation of mass balances, many model parameters are
defined by first principles. However, some parameters, namely the
observation and state covariance matrices, need to be estimated
from process data before the dynamic Bayesian methods could be
applied. This thesis makes use of Bayesian machine learning
techniques to estimate these parameters, separating process
disturbances from instrument measurement noise.
Subjects/Keywords: Gross Error Detection; Bayesian Inference; Data Reconciliation
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Gonzalez, R. (2010). Bayesian Methods for On-Line Gross Error Detection and
Compensation. (Masters Thesis). University of Alberta. Retrieved from https://era.library.ualberta.ca/files/w6634487q
Chicago Manual of Style (16th Edition):
Gonzalez, Ruben. “Bayesian Methods for On-Line Gross Error Detection and
Compensation.” 2010. Masters Thesis, University of Alberta. Accessed March 04, 2021.
https://era.library.ualberta.ca/files/w6634487q.
MLA Handbook (7th Edition):
Gonzalez, Ruben. “Bayesian Methods for On-Line Gross Error Detection and
Compensation.” 2010. Web. 04 Mar 2021.
Vancouver:
Gonzalez R. Bayesian Methods for On-Line Gross Error Detection and
Compensation. [Internet] [Masters thesis]. University of Alberta; 2010. [cited 2021 Mar 04].
Available from: https://era.library.ualberta.ca/files/w6634487q.
Council of Science Editors:
Gonzalez R. Bayesian Methods for On-Line Gross Error Detection and
Compensation. [Masters Thesis]. University of Alberta; 2010. Available from: https://era.library.ualberta.ca/files/w6634487q
Cornell University
15.
Wan, Muting.
Model-Based Classification With Applications To High-Dimensional Data In Bioinformatics.
Degree: PhD, Statistics, 2015, Cornell University
URL: http://hdl.handle.net/1813/39389
► In recent years, sparse classification problems have emerged in many fields of study. Finite mixture models have been developed to facilitate Bayesian inference where parameter…
(more)
▼ In recent years, sparse classification problems have emerged in many fields of study. Finite mixture models have been developed to facilitate
Bayesian inference where parameter sparsity is substantial. Shrinkage estimation allows strength borrowing across features in light of the parallel nature of multiple hypothesis tests. Important examples that incorporate shrinkage estimation and finite mixture model for sparse classification include the hierarchical model in Smyth (2004) and the explicit mixture model in Bar et al. (2010) for
Bayesian microarray analysis. Classification with finite mixture models is based on the posterior expectation of latent indicator variables. These quantities are typically estimated using the expectation-maximization (EM) algorithm in an empirical Bayes approach or Markov chain Monte Carlo (MCMC) in a fully
Bayesian approach. MCMC is limited in applicability where high-dimensional data are involved because its sampling-based nature leads to slow computations and hard-to-monitor convergence. In a fully
Bayesian framework, we investigate the feasibility and performance of variational Bayes (VB) approximation and apply the VB approach to fully
Bayesian versions of several finite mixture models that have been proposed in bioinformatics. We find that it achieves desirable speed and accuracy in sparse classification with hierarchical mixture models for high-dimensional data. Another example of sparse classification in bioinformatics solvable via model-based approaches is expression quantitative trait loci (eQTL) detection, in which determining whether association between a gene and any given single nucleotide polymorphism (SNP) is significant is regarded as classifying genes as null or non-null with respect to the given SNP. High-dimensionality of the data not only causes difficulties in computations, but also renders the confounding impact of unwanted variation in the data irrefutable. Model-based approaches that account for unwanted variation by incorporating a factor analysis term representing hidden factors and their effects have been adopted in applications such as differential analysis and eQTL detection. HEFT (Gao et al., 2014) is a fast approach for model-based eQTL identification while simultaneously learning hidden effects. We develop a hierarchical mixture model-based empirical Bayes approach for sparse classification while simultaneously accounting for unwanted variation, as well as a family of model-based approaches that are its simplifications with the aim of attractive computational efficiency. We investigate feasibility and performance of these model-based approaches in comparison with HEFT using several real data examples in bioinformatics.
Advisors/Committee Members: Booth, James (chair), Hooker, Giles J. (committee member), Wells, Martin Timothy (committee member).
Subjects/Keywords: Bayesian inference; Linear mixed models; Bioinformatics
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Wan, M. (2015). Model-Based Classification With Applications To High-Dimensional Data In Bioinformatics. (Doctoral Dissertation). Cornell University. Retrieved from http://hdl.handle.net/1813/39389
Chicago Manual of Style (16th Edition):
Wan, Muting. “Model-Based Classification With Applications To High-Dimensional Data In Bioinformatics.” 2015. Doctoral Dissertation, Cornell University. Accessed March 04, 2021.
http://hdl.handle.net/1813/39389.
MLA Handbook (7th Edition):
Wan, Muting. “Model-Based Classification With Applications To High-Dimensional Data In Bioinformatics.” 2015. Web. 04 Mar 2021.
Vancouver:
Wan M. Model-Based Classification With Applications To High-Dimensional Data In Bioinformatics. [Internet] [Doctoral dissertation]. Cornell University; 2015. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/1813/39389.
Council of Science Editors:
Wan M. Model-Based Classification With Applications To High-Dimensional Data In Bioinformatics. [Doctoral Dissertation]. Cornell University; 2015. Available from: http://hdl.handle.net/1813/39389
Penn State University
16.
Schall, Megan Victoria.
Comparative Bioenergetics of Two Lake Trout Morphotypes.
Degree: 2013, Penn State University
URL: https://submit-etda.libraries.psu.edu/catalog/18506
► Lake trout (Salvelinus namaycush) are deep water apex predators native to North America that inhabit glacially formed lakes, including the Laurentian Great Lakes. Lake trout…
(more)
▼ Lake trout (Salvelinus namaycush) are deep water apex predators native to North America that inhabit glacially formed lakes, including the Laurentian Great Lakes. Lake trout populations in the Great Lakes are recovering from overfishing, sea lamprey parasitism, and poor water quality conditions that occurred during the 19th and 20th centuries, in addition to present day stressors (e.g., invasive species, climate change, etc.). Current stocking and management strategies aim to produce self-sustaining and viable populations. Management is complicated, however, by the presence of multiple body forms (i.e., morphotypes) of lake trout which differ in habitat utilization, prey consumption, lipid storage, and spawning preferences. Bioenergetics models are useful management tools that relate fish physiology to habitat usage (e.g., prey consumed or growth predictions at a given temperature). However, there is currently only a single bioenergetics model developed for one lake trout morphotype (lean), and this model failed to incorporate temperatures outside lake trout’s preferred range. To investigate potential morphotype differences and develop bioenergetics models with broader temperature applications, I completed consumption and respiration experiments on two actively stocked lake trout morphotypes, lean and humper, across a wide range of temperatures (4-22°C) and size classes (5-100g).
Bayesian estimation was used during model development to propagate uncertainty through to final growth predictions. Morphotype differences were minimal, but when present, were temperature and weight dependent. Basal respiration did not differ among the morphotypes at any temperature or size class tested. Growth and consumption differences were subtle and not consistent across size ranges tested. Management scenarios investigated (e.g., predicted growth at an average temperature of 11.7°C and 14.4°C over a 30 day period) demonstrated no differences in growth between the two morphotypes utilizing temperatures presently found in the Great Lakes. Yet, management decisions should not be based on energetics alone as other factors including available habitat, prey resources, and competition should also be considered. Managers should consider the characteristics of both the population structure and habitat conditions when developing recommendations for management of lake trout morphotypes.
Advisors/Committee Members: Tyler Wagner, Thesis Advisor/Co-Advisor.
Subjects/Keywords: bioenergetics modeling; lake trout; Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Schall, M. V. (2013). Comparative Bioenergetics of Two Lake Trout Morphotypes. (Thesis). Penn State University. Retrieved from https://submit-etda.libraries.psu.edu/catalog/18506
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Schall, Megan Victoria. “Comparative Bioenergetics of Two Lake Trout Morphotypes.” 2013. Thesis, Penn State University. Accessed March 04, 2021.
https://submit-etda.libraries.psu.edu/catalog/18506.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Schall, Megan Victoria. “Comparative Bioenergetics of Two Lake Trout Morphotypes.” 2013. Web. 04 Mar 2021.
Vancouver:
Schall MV. Comparative Bioenergetics of Two Lake Trout Morphotypes. [Internet] [Thesis]. Penn State University; 2013. [cited 2021 Mar 04].
Available from: https://submit-etda.libraries.psu.edu/catalog/18506.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Schall MV. Comparative Bioenergetics of Two Lake Trout Morphotypes. [Thesis]. Penn State University; 2013. Available from: https://submit-etda.libraries.psu.edu/catalog/18506
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
North Carolina State University
17.
DiCasoli, Carl Matthew.
Bayesian Regression Methods for Crossing Survival Curves.
Degree: PhD, Statistics, 2009, North Carolina State University
URL: http://www.lib.ncsu.edu/resolver/1840.16/4743
► In survival data analysis, the proportional hazards (PH), accelerated failure time (AFT), and proportional odds (PO) models are commonly used semiparametric models for the comparison…
(more)
▼ In survival data analysis, the proportional hazards (PH), accelerated failure time
(AFT), and proportional odds (PO) models are commonly used semiparametric models for
the comparison of survivability in subjects. These models assume that the survival curves
do not cross. However, in some clinical applications, the survival curves pertaining to the two groups of subjects under the study may cross each other, especially for long-duration
studies. Hence, these three models stated above may no longer be suitable for making
inference
Yang and Prentice (2005) proposed a model which separately models the short-term and long-term hazard ratios nesting both PH and PO. This feature allows for the survival functions to cross. First, we study the estimation procedure in the Yang-Prentice model with regards to the two-sample case. We propose two different approaches: (1)
Bayesian bootstrap and (2) smoothing methods. The first approach involves
Bayesian bootstrap
with likelihoods corresponding to binomial and Poisson forms while the second approach
involves kernel smoothing methods as well as smoothing spline methods. A simulation is
conducted to compare various methods under the two-sample case. Next, we extend the
Yang-Prentice model to a regression version involving predictors and examine three likelihood
approaches including Poisson form, pseudo-likelihood, and
Bayesian smoothing. The
effects of model misspecification on asymptotic relative efficiency are also studied empirically. The results from simulation studies indicate that the PH, AFT, and PO models are not robust to model misspecifications when the survival functions are allowed to cross.
Finally, we calculate the marginal density via variational methods to determine
the Bayes factor. Either a full
Bayesian or
Bayesian approach is implemented to perform
model selection. Both approaches accurately identify the correct model, even under slight
misspecification, and are computationally more efficient than MCMC techniques.
Advisors/Committee Members: Dr. Charles Apperson, Committee Member (advisor), Dr. Wenbin Lu, Committee Member (advisor), Dr. Brian Reich, Committee Member (advisor), Dr. Subhashis Ghosal, Committee Co-Chair (advisor), Dr. Sujit Ghosh, Committee Chair (advisor).
Subjects/Keywords: variational methods; Bayesian inference; survival analysis
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
DiCasoli, C. M. (2009). Bayesian Regression Methods for Crossing Survival Curves. (Doctoral Dissertation). North Carolina State University. Retrieved from http://www.lib.ncsu.edu/resolver/1840.16/4743
Chicago Manual of Style (16th Edition):
DiCasoli, Carl Matthew. “Bayesian Regression Methods for Crossing Survival Curves.” 2009. Doctoral Dissertation, North Carolina State University. Accessed March 04, 2021.
http://www.lib.ncsu.edu/resolver/1840.16/4743.
MLA Handbook (7th Edition):
DiCasoli, Carl Matthew. “Bayesian Regression Methods for Crossing Survival Curves.” 2009. Web. 04 Mar 2021.
Vancouver:
DiCasoli CM. Bayesian Regression Methods for Crossing Survival Curves. [Internet] [Doctoral dissertation]. North Carolina State University; 2009. [cited 2021 Mar 04].
Available from: http://www.lib.ncsu.edu/resolver/1840.16/4743.
Council of Science Editors:
DiCasoli CM. Bayesian Regression Methods for Crossing Survival Curves. [Doctoral Dissertation]. North Carolina State University; 2009. Available from: http://www.lib.ncsu.edu/resolver/1840.16/4743
18.
Feldman, Naomi H.
Interactions between word and speech sound categorization in
language acquisition.
Degree: PhD, Cognitive Sciences, 2011, Brown University
URL: https://repository.library.brown.edu/studio/item/bdr:11219/
► Infants learn to segment words from fluent speech during the same period as they learn native language phonetic categories, yet accounts of phonetic category acquisition…
(more)
▼ Infants learn to segment words from fluent speech
during the same period as they learn native language phonetic
categories, yet accounts of phonetic category acquisition typically
ignore information about the words in which sounds appear. This
work uses computational and behavioral methods to explore the
hypothesis that the words infants segment from fluent speech can
provide useful cues to guide phonetic category acquisition. An
interactive
Bayesian learning model is created to illustrate how
feedback from segmented words might constrain phonetic category
learning by providing information about which sounds occur together
in words. This model is compared with two models that learn
phonetic categories only from distributions of sounds in acoustic
space. Simulations show that word-level information can
successfully disambiguate overlapping English vowel categories,
leading to more robust category learning than distributional
information alone. Next, two experiments test whether human
learners can make use of the word-level cues required for this type
of interactive learning. These experiments demonstrate that adult
learners are sensitive to cooccurrence patterns of sounds in
acoustic word tokens of an unfamiliar language. However, human
learners appear to treat the patterns differently when words are
heard in isolation versus when they are heard in fluent speech,
adopting either a word-level or a phonological interpretation
depending on experimental context. These behavioral results
partially support the predictions of the model but underscore the
complexity of the phonetic category learning problem. Together, the
computational and behavioral results suggest that phonetic category
learning can be better understood in conjunction with other
contemporaneous learning processes and that simultaneous learning
of multiple layers of linguistic structure can potentially make the
language acquisition problem more tractable.
Advisors/Committee Members: Morgan, James (Director), Griffiths, Thomas (Reader), Blumstein, Sheila (Reader).
Subjects/Keywords: language acquisition; phonetic category learning; Bayesian
inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Feldman, N. H. (2011). Interactions between word and speech sound categorization in
language acquisition. (Doctoral Dissertation). Brown University. Retrieved from https://repository.library.brown.edu/studio/item/bdr:11219/
Chicago Manual of Style (16th Edition):
Feldman, Naomi H. “Interactions between word and speech sound categorization in
language acquisition.” 2011. Doctoral Dissertation, Brown University. Accessed March 04, 2021.
https://repository.library.brown.edu/studio/item/bdr:11219/.
MLA Handbook (7th Edition):
Feldman, Naomi H. “Interactions between word and speech sound categorization in
language acquisition.” 2011. Web. 04 Mar 2021.
Vancouver:
Feldman NH. Interactions between word and speech sound categorization in
language acquisition. [Internet] [Doctoral dissertation]. Brown University; 2011. [cited 2021 Mar 04].
Available from: https://repository.library.brown.edu/studio/item/bdr:11219/.
Council of Science Editors:
Feldman NH. Interactions between word and speech sound categorization in
language acquisition. [Doctoral Dissertation]. Brown University; 2011. Available from: https://repository.library.brown.edu/studio/item/bdr:11219/
Tampere University
19.
Lu, Chien.
An Improved Nearest Neighbor Based Entropy Estimator with Local Ellipsoid Correction and its Application to Evaluation of MCMC Posterior Samples
.
Degree: 2018, Tampere University
URL: https://trepo.tuni.fi/handle/10024/104066
► Entropy estimation is an important technique to summarize the uncertainty of a distribution underlying a set of samples. It ties to important research problems in…
(more)
▼ Entropy estimation is an important technique to summarize the uncertainty of a distribution underlying a set of samples. It ties to important research problems in fields such as statistics, machine learning and so on. The k-nearest neighbor (kNN) estimator is one widely used classical nonparametric method although it suffers bias issue especially when the dimensionality of the data is high.
In this thesis, an improved kNN entropy estimator is developed. The proposed method has the advantage of a learning a local ellipsoid to be used in the estimation, in order to mitigate the bias issue which results from the local uniformity. Several numerical experiments have been conducted and the results have shown that the proposed approach can efficiently reduce the bias especially in when the dimension is high.
Another studied topic in this thesis is the evaluation of the correctness of the posterior samples when conducting Bayesian inferences. This thesis demonstrates that the proposed estimator can be applied to such a task. We show that the simulation-based approach is more efficient and discriminative than a lower bound based method by one simple experiment, and the proposed kNN estimation can improve the accuracy of the state-of-the-art simulation-based approach.
Subjects/Keywords: entropy estimation;
nonparametric estimator;
Bayesian inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Lu, C. (2018). An Improved Nearest Neighbor Based Entropy Estimator with Local Ellipsoid Correction and its Application to Evaluation of MCMC Posterior Samples
. (Masters Thesis). Tampere University. Retrieved from https://trepo.tuni.fi/handle/10024/104066
Chicago Manual of Style (16th Edition):
Lu, Chien. “An Improved Nearest Neighbor Based Entropy Estimator with Local Ellipsoid Correction and its Application to Evaluation of MCMC Posterior Samples
.” 2018. Masters Thesis, Tampere University. Accessed March 04, 2021.
https://trepo.tuni.fi/handle/10024/104066.
MLA Handbook (7th Edition):
Lu, Chien. “An Improved Nearest Neighbor Based Entropy Estimator with Local Ellipsoid Correction and its Application to Evaluation of MCMC Posterior Samples
.” 2018. Web. 04 Mar 2021.
Vancouver:
Lu C. An Improved Nearest Neighbor Based Entropy Estimator with Local Ellipsoid Correction and its Application to Evaluation of MCMC Posterior Samples
. [Internet] [Masters thesis]. Tampere University; 2018. [cited 2021 Mar 04].
Available from: https://trepo.tuni.fi/handle/10024/104066.
Council of Science Editors:
Lu C. An Improved Nearest Neighbor Based Entropy Estimator with Local Ellipsoid Correction and its Application to Evaluation of MCMC Posterior Samples
. [Masters Thesis]. Tampere University; 2018. Available from: https://trepo.tuni.fi/handle/10024/104066
Baylor University
20.
Casement, Christopher James, 1987-.
Graphical methods in prior elicitation.
Degree: PhD, Baylor University. Dept. of Statistical Sciences., 2017, Baylor University
URL: http://hdl.handle.net/2104/10111
► Prior elicitation is the process of quantifying an expert's belief in the form of a probability distribution on a parameter(s) to be used in a…
(more)
▼ Prior elicitation is the process of quantifying an expert's belief in the form of a probability distribution on a parameter(s) to be used in a
Bayesian data analysis. Existing methods require experts to quantify their belief by specifying multiple distribution summaries, which are then converted into the parameters of a given prior family. The resulting priors, however, may not accurately represent the expert's opinion, which in turn can undermine the accuracy of an analysis. In this dissertation we propose two interactive graphical strategies for prior elicitation, along with web-based Shiny implementations for each, that do not rely on an expert's ability to reliably quantify thier beliefs. Instead, the expert moves through a series of tests where they are tasked with selecting hypothetical future datasets they believe to be most likely from a collection of candidate datasets that are presented in graphical form. The algorithms then convert these selections into a prior distribution on the parameter(s) of interest. After discussing each elicitation method, we propose a variation on the Metropolis-Hastings algorithm that provides support for the underlying stochastic scheme in the second of the two elicitation strategies. We apply the methods to data models that are commonly employed in practice, such as Bernoulli, Poisson, and Normal, though the methods can be more generally applied to other univariate data models.
Advisors/Committee Members: Kahle, David J. (advisor).
Subjects/Keywords: Bayesian statistics. Prior elicitation. Graphical inference.
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Casement, Christopher James, 1. (2017). Graphical methods in prior elicitation. (Doctoral Dissertation). Baylor University. Retrieved from http://hdl.handle.net/2104/10111
Chicago Manual of Style (16th Edition):
Casement, Christopher James, 1987-. “Graphical methods in prior elicitation.” 2017. Doctoral Dissertation, Baylor University. Accessed March 04, 2021.
http://hdl.handle.net/2104/10111.
MLA Handbook (7th Edition):
Casement, Christopher James, 1987-. “Graphical methods in prior elicitation.” 2017. Web. 04 Mar 2021.
Vancouver:
Casement, Christopher James 1. Graphical methods in prior elicitation. [Internet] [Doctoral dissertation]. Baylor University; 2017. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2104/10111.
Council of Science Editors:
Casement, Christopher James 1. Graphical methods in prior elicitation. [Doctoral Dissertation]. Baylor University; 2017. Available from: http://hdl.handle.net/2104/10111
University of Houston
21.
Bhardwaj, Manisha 1986-.
Visual decision making in the presence of stimulus and measurement correlations.
Degree: PhD, Mathematics, 2013, University of Houston
URL: http://hdl.handle.net/10657/1239
► Our brains process sensory information to infer the state of the world. However, the input from our senses is noisy, which may lead to errors…
(more)
▼ Our brains process sensory information to infer the state of the world. However, the input from our senses is noisy, which may lead to errors in perceptual judgements. A number of theoretical studies have modeled perception as a process of probabilistic
inference that involves making decisions based on uncertain evidence.
Bayesian optimality is a general principle of probabilistic
inference that has been successfully used to build quantitative models of perception. In addition, several experimental studies show that human observers make best possible decisions, and hence exhibit close to Bayes-optimal behavior on various visual perceptual tasks such as visual search, sameness judgement, and change detection. However, the impact of structured stimuli on decision-making remains largely unexplored. Moreover, the sensory measurements can themselves be strongly correlated to produce a structured representation of the stimulus input. These measurement correlations can interact with the structure of the external input in many possible ways and should not be considered in isolation.
In this work, we focus on visual search task to examine how visual perception is affected by structured input. We analyze the responses of subjects on a target detection experiment where the stimulus orientations were generated with varying strength of correlations across different experimental sessions. We fit several models to the experimental data using maximum-likelihood parameter estimation. We use rigorous model selection to find that human observers take into account stimulus correlations in detecting a target. However, they behave suboptimally in inferring the correct stimulus correlations that were used in the experiment. We find that perhaps observers treat the partial stimulus correlations identically and behave differently when the stimuli are perfectly correlated.
We also describe how the relation between measurement and stimulus correlations affects the performance of an ideal
Bayesian observer in a family of target detection tasks. We find that the effect of measurement correlations depends on its interaction with stimulus correlations and other statistical structure parameters. Measurement correlations always improves the performance of the ideal observer on a detection task with multiple targets; whereas in the case of single target, the impact is significant only in the presence of strong external structure.
Advisors/Committee Members: Josić, Krešimir (advisor), Nicol, Matthew (committee member), Timofeyev, Ilya (committee member), Ma, Wei Ji (committee member).
Subjects/Keywords: Bayesian inference; Model comparison; Correlations; Target detection
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Bhardwaj, M. 1. (2013). Visual decision making in the presence of stimulus and measurement correlations. (Doctoral Dissertation). University of Houston. Retrieved from http://hdl.handle.net/10657/1239
Chicago Manual of Style (16th Edition):
Bhardwaj, Manisha 1986-. “Visual decision making in the presence of stimulus and measurement correlations.” 2013. Doctoral Dissertation, University of Houston. Accessed March 04, 2021.
http://hdl.handle.net/10657/1239.
MLA Handbook (7th Edition):
Bhardwaj, Manisha 1986-. “Visual decision making in the presence of stimulus and measurement correlations.” 2013. Web. 04 Mar 2021.
Vancouver:
Bhardwaj M1. Visual decision making in the presence of stimulus and measurement correlations. [Internet] [Doctoral dissertation]. University of Houston; 2013. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/10657/1239.
Council of Science Editors:
Bhardwaj M1. Visual decision making in the presence of stimulus and measurement correlations. [Doctoral Dissertation]. University of Houston; 2013. Available from: http://hdl.handle.net/10657/1239
University of Cambridge
22.
Graff, Philip B.
Bayesian methods for gravitational waves and neural networks.
Degree: PhD, 2012, University of Cambridge
URL: https://doi.org/10.17863/CAM.16592
;
https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.566190
► Einstein’s general theory of relativity has withstood 100 years of testing and will soon be facing one of its toughest challenges. In a few years…
(more)
▼ Einstein’s general theory of relativity has withstood 100 years of testing and will soon be facing one of its toughest challenges. In a few years we expect to be entering the era of the first direct observations of gravitational waves. These are tiny perturbations of space-time that are generated by accelerating matter and affect the measured distances between two points. Observations of these using the laser interferometers, which are the most sensitive length-measuring devices in the world, will allow us to test models of interactions in the strong field regime of gravity and eventually general relativity itself. I apply the tools of Bayesian inference for the examination of gravitational wave data from the LIGO and Virgo detectors. This is used for signal detection and estimation of the source parameters. I quantify the ability of a network of ground-based detectors to localise a source position on the sky for electromagnetic follow-up. Bayesian criteria are also applied to separating real signals from glitches in the detectors. These same tools and lessons can also be applied to the type of data expected from planned space-based detectors. Using simulations from the Mock LISA Data Challenges, I analyse our ability to detect and characterise both burst and continuous signals. The two seemingly different signal types will be overlapping and confused with one another for a space-based detector; my analysis shows that we will be able to separate and identify many signals present. Data sets and astrophysical models are continuously increasing in complexity. This will create an additional computational burden for performing Bayesian inference and other types of data analysis. I investigate the application of the MOPED algorithm for faster parameter estimation and data compression. I find that its shortcomings make it a less favourable candidate for further implementation. The framework of an artificial neural network is a simple model for the structure of a brain which can “learn” functional relationships between sets of inputs and outputs. I describe an algorithm developed for the training of feed-forward networks on pre-calculated data sets. The trained networks can then be used for fast prediction of outputs for new sets of inputs. After demonstrating capabilities on toy data sets, I apply the ability of the network to classifying handwritten digits from the MNIST database and measuring ellipticities of galaxies in the Mapping Dark Matter challenge. The power of neural networks for learning and rapid prediction is also useful in Bayesian inference where the likelihood function is computationally expensive. The new BAMBI algorithm is detailed, in which our network training algorithm is combined with the nested sampling algorithm MULTINEST to provide rapid Bayesian inference. Using samples from the normal inference, a network is trained on the likelihood function and eventually used in its place. This is able to provide significant increase in the speed of Bayesian inference while returning identical results. The…
Subjects/Keywords: 530; Bayesian inference; Gravitational waves; Machine learning
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Graff, P. B. (2012). Bayesian methods for gravitational waves and neural networks. (Doctoral Dissertation). University of Cambridge. Retrieved from https://doi.org/10.17863/CAM.16592 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.566190
Chicago Manual of Style (16th Edition):
Graff, Philip B. “Bayesian methods for gravitational waves and neural networks.” 2012. Doctoral Dissertation, University of Cambridge. Accessed March 04, 2021.
https://doi.org/10.17863/CAM.16592 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.566190.
MLA Handbook (7th Edition):
Graff, Philip B. “Bayesian methods for gravitational waves and neural networks.” 2012. Web. 04 Mar 2021.
Vancouver:
Graff PB. Bayesian methods for gravitational waves and neural networks. [Internet] [Doctoral dissertation]. University of Cambridge; 2012. [cited 2021 Mar 04].
Available from: https://doi.org/10.17863/CAM.16592 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.566190.
Council of Science Editors:
Graff PB. Bayesian methods for gravitational waves and neural networks. [Doctoral Dissertation]. University of Cambridge; 2012. Available from: https://doi.org/10.17863/CAM.16592 ; https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.566190
University of Cambridge
23.
Jin, Yingyan.
Supervised learning for back analysis of excavations in the observational method.
Degree: PhD, 2018, University of Cambridge
URL: https://www.repository.cam.ac.uk/handle/1810/275587
► In the past few decades, demand for construction in underground spaces has increased dramatically in urban areas with high population densities. However, the impact of…
(more)
▼ In the past few decades, demand for construction in underground spaces has increased dramatically in urban areas with high population densities. However, the impact of the construction of underground structures on surrounding infrastructure raises concerns since movements caused by deep excavations might damage adjacent buildings. Unfortunately, the prediction of geotechnical behaviour is difficult due to uncertainties and lack of information of on the underground environment. Therefore, to ensure safety, engineers tend to choose very conservative designs that result in requiring unnecessary material and longer construction time. The observational method, which was proposed by Peck in 1969, and formalised in Eurocode 7 in 1987, provides a way to avoid such redundancy by modifying the design based on the knowledge gathered during construction. The review process within the observational method is recognised as back analysis.
Supervised learning can aid in this process, providing a systematic procedure to assess soil parameters based on monitoring data and prediction of the ground response. A probabilistic model is developed in this research to account for the uncertainties in the problem. Sequential Bayesian inference is used to update the soil parameters at each excavation stage when observations are available. The accuracy of the prediction for future stages improves at each stage. Meanwhile, the uncertainty contained in the prediction decreases, and therefore the confidence on the corresponding design also increases. Moreover, the Bayesian method integrates subjective engineering experience and objective observations in a rational and quantitative way, which enables the model to update soil parameters even when the amount of data is very limited. It also allows the use of the knowledge learnt from comparable ground conditions, which is particularly useful in the absence of site-specific information on ground conditions.
Four probabilistic models are developed in this research. The first two incorporate empirical excavation design methods. These simple models are used to examine the practicality of the approach with several cases. The next two are coupled with a program called FREW, which is able to simulate the excavation process, still in a relatively simplistic way. The baseline model with simple assumptions on model error and another one is a more sophisticated model considering measurement error and spatial relationships among the observations. Their efficiency and accuracy are verified using a synthetic case and tested based on a case history from the London Crossrail project. In the end, the models are compared and their flexibility in different cases is discussed.
Subjects/Keywords: Bayesian inference; Excavations; Observational method; Back analysis
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Jin, Y. (2018). Supervised learning for back analysis of excavations in the observational method. (Doctoral Dissertation). University of Cambridge. Retrieved from https://www.repository.cam.ac.uk/handle/1810/275587
Chicago Manual of Style (16th Edition):
Jin, Yingyan. “Supervised learning for back analysis of excavations in the observational method.” 2018. Doctoral Dissertation, University of Cambridge. Accessed March 04, 2021.
https://www.repository.cam.ac.uk/handle/1810/275587.
MLA Handbook (7th Edition):
Jin, Yingyan. “Supervised learning for back analysis of excavations in the observational method.” 2018. Web. 04 Mar 2021.
Vancouver:
Jin Y. Supervised learning for back analysis of excavations in the observational method. [Internet] [Doctoral dissertation]. University of Cambridge; 2018. [cited 2021 Mar 04].
Available from: https://www.repository.cam.ac.uk/handle/1810/275587.
Council of Science Editors:
Jin Y. Supervised learning for back analysis of excavations in the observational method. [Doctoral Dissertation]. University of Cambridge; 2018. Available from: https://www.repository.cam.ac.uk/handle/1810/275587
24.
Graff, Philip B.
Bayesian methods for gravitational waves and neural networks.
Degree: PhD, 2012, University of Cambridge
URL: http://www.dspace.cam.ac.uk/handle/1810/244270https://www.repository.cam.ac.uk/bitstream/1810/244270/2/license.txt
;
https://www.repository.cam.ac.uk/bitstream/1810/244270/3/license_url
;
https://www.repository.cam.ac.uk/bitstream/1810/244270/4/license_text
;
https://www.repository.cam.ac.uk/bitstream/1810/244270/5/license_rdf
;
https://www.repository.cam.ac.uk/bitstream/1810/244270/8/thesis.pdf.txt
;
https://www.repository.cam.ac.uk/bitstream/1810/244270/9/thesis.pdf.jpg
► Einstein’s general theory of relativity has withstood 100 years of testing and will soon be facing one of its toughest challenges. In a few years…
(more)
▼ Einstein’s general theory of relativity has withstood 100 years of testing
and will soon be facing one of its toughest challenges. In a few years
we expect to be entering the era of the first direct observations of gravitational
waves. These are tiny perturbations of space-time that are generated
by accelerating matter and affect the measured distances between
two points. Observations of these using the laser interferometers, which
are the most sensitive length-measuring devices in the world, will allow
us to test models of interactions in the strong field regime of gravity and
eventually general relativity itself.
I apply the tools of Bayesian inference for the examination of gravitational
wave data from the LIGO and Virgo detectors. This is used for signal
detection and estimation of the source parameters. I quantify the ability
of a network of ground-based detectors to localise a source position
on the sky for electromagnetic follow-up. Bayesian criteria are also applied
to separating real signals from glitches in the detectors. These same
tools and lessons can also be applied to the type of data expected from
planned space-based detectors. Using simulations from the Mock LISA
Data Challenges, I analyse our ability to detect and characterise both burst
and continuous signals. The two seemingly different signal types will be
overlapping and confused with one another for a space-based detector; my
analysis shows that we will be able to separate and identify many signals
present.
Data sets and astrophysical models are continuously increasing in complexity.
This will create an additional computational burden for performing
Bayesian inference and other types of data analysis. I investigate the
application of the MOPED algorithm for faster parameter estimation and
data compression. I find that its shortcomings make it a less favourable
candidate for further implementation.
The framework of an artificial neural network is a simple model for the
structure of a brain which can “learn” functional relationships between sets
of inputs and outputs. I describe an algorithm developed for the training of
feed-forward networks on pre-calculated data sets. The trained networks
can then be used for fast prediction of outputs for new sets of inputs. After
demonstrating capabilities on toy data sets, I apply the ability of the
network to classifying handwritten digits from the MNIST database and
measuring ellipticities of galaxies in the Mapping Dark Matter challenge.
The power of neural networks for learning and rapid prediction is also
useful in Bayesian inference where the likelihood function is computationally
expensive. The new BAMBI algorithm is detailed, in which our
network training algorithm is combined with the nested sampling algorithm
MULTINEST to provide rapid Bayesian inference. Using samples
from the normal inference, a network is trained on the likelihood function
and eventually used in its place. This is able to provide significant increase
in the speed of Bayesian inference…
Subjects/Keywords: Bayesian inference; Gravitational waves; Machine learning
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Graff, P. B. (2012). Bayesian methods for gravitational waves and neural networks. (Doctoral Dissertation). University of Cambridge. Retrieved from http://www.dspace.cam.ac.uk/handle/1810/244270https://www.repository.cam.ac.uk/bitstream/1810/244270/2/license.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/3/license_url ; https://www.repository.cam.ac.uk/bitstream/1810/244270/4/license_text ; https://www.repository.cam.ac.uk/bitstream/1810/244270/5/license_rdf ; https://www.repository.cam.ac.uk/bitstream/1810/244270/8/thesis.pdf.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/9/thesis.pdf.jpg
Chicago Manual of Style (16th Edition):
Graff, Philip B. “Bayesian methods for gravitational waves and neural networks.” 2012. Doctoral Dissertation, University of Cambridge. Accessed March 04, 2021.
http://www.dspace.cam.ac.uk/handle/1810/244270https://www.repository.cam.ac.uk/bitstream/1810/244270/2/license.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/3/license_url ; https://www.repository.cam.ac.uk/bitstream/1810/244270/4/license_text ; https://www.repository.cam.ac.uk/bitstream/1810/244270/5/license_rdf ; https://www.repository.cam.ac.uk/bitstream/1810/244270/8/thesis.pdf.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/9/thesis.pdf.jpg.
MLA Handbook (7th Edition):
Graff, Philip B. “Bayesian methods for gravitational waves and neural networks.” 2012. Web. 04 Mar 2021.
Vancouver:
Graff PB. Bayesian methods for gravitational waves and neural networks. [Internet] [Doctoral dissertation]. University of Cambridge; 2012. [cited 2021 Mar 04].
Available from: http://www.dspace.cam.ac.uk/handle/1810/244270https://www.repository.cam.ac.uk/bitstream/1810/244270/2/license.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/3/license_url ; https://www.repository.cam.ac.uk/bitstream/1810/244270/4/license_text ; https://www.repository.cam.ac.uk/bitstream/1810/244270/5/license_rdf ; https://www.repository.cam.ac.uk/bitstream/1810/244270/8/thesis.pdf.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/9/thesis.pdf.jpg.
Council of Science Editors:
Graff PB. Bayesian methods for gravitational waves and neural networks. [Doctoral Dissertation]. University of Cambridge; 2012. Available from: http://www.dspace.cam.ac.uk/handle/1810/244270https://www.repository.cam.ac.uk/bitstream/1810/244270/2/license.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/3/license_url ; https://www.repository.cam.ac.uk/bitstream/1810/244270/4/license_text ; https://www.repository.cam.ac.uk/bitstream/1810/244270/5/license_rdf ; https://www.repository.cam.ac.uk/bitstream/1810/244270/8/thesis.pdf.txt ; https://www.repository.cam.ac.uk/bitstream/1810/244270/9/thesis.pdf.jpg
University of Adelaide
25.
Walker, James.
Bayesian Inference and Model Selection for Partially-Observed, Continuous-Time, Stochastic Epidemic Models.
Degree: 2019, University of Adelaide
URL: http://hdl.handle.net/2440/124703
► Emerging infectious diseases are an ongoing threat to the health of populations around the world. In response, countries such as the USA, UK and Australia,…
(more)
▼ Emerging infectious diseases are an ongoing threat to the health of populations around the world. In response, countries such as the USA, UK and Australia, have outlined data collection protocols to surveil these novel diseases. One of the aims of these data collection protocols is to characterise the disease in terms of transmissibility and clinical severity in order to inform an appropriate public health response. This kind of data collection protocol is yet to be enacted in Australia, but such a protocol is likely to be tested during a seasonal in uenza ( u) outbreak in the next few years. However, it is important that methods for characterising these diseases are ready and well understood for when an epidemic disease emerges. The epidemic may only be characterised well if its dynamics are well described (by a model) and are accurately quanti ed (by precisely inferred model parameters). This thesis models epidemics and the data collection process as partially-observed continuous-time Markov chains and aims to choose between models and infer parameters using early outbreak data. It develops
Bayesian methods to infer epidemic parameters from data on multiple small outbreaks, and outbreaks in a population of households. An exploratory analysis is conducted to assess the accuracy and precision of parameter estimates under di erent epidemic surveillance schemes, di erent models and di erent kinds of model misspeci cation. It describes a novel
Bayesian model selection method and employs it to infer two important characteristics for understanding emerging epidemics: the shape of the infectious period distribution; and, the time of infectiousness relative to symptom onset. Lastly, this thesis outlines a method for jointly inferring model parameters and selecting between epidemic models. This new method is compared with an existing method on two epidemic models and is applied to a di cult model selection problem.
Advisors/Committee Members: Ross, Joshua (advisor), Black, Andrew (advisor), School of Mathematical Sciences (school).
Subjects/Keywords: Epidemiology; Bayesian Inference; Model Selection; Stochastic Modelling
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Walker, J. (2019). Bayesian Inference and Model Selection for Partially-Observed, Continuous-Time, Stochastic Epidemic Models. (Thesis). University of Adelaide. Retrieved from http://hdl.handle.net/2440/124703
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Walker, James. “Bayesian Inference and Model Selection for Partially-Observed, Continuous-Time, Stochastic Epidemic Models.” 2019. Thesis, University of Adelaide. Accessed March 04, 2021.
http://hdl.handle.net/2440/124703.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Walker, James. “Bayesian Inference and Model Selection for Partially-Observed, Continuous-Time, Stochastic Epidemic Models.” 2019. Web. 04 Mar 2021.
Vancouver:
Walker J. Bayesian Inference and Model Selection for Partially-Observed, Continuous-Time, Stochastic Epidemic Models. [Internet] [Thesis]. University of Adelaide; 2019. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2440/124703.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Walker J. Bayesian Inference and Model Selection for Partially-Observed, Continuous-Time, Stochastic Epidemic Models. [Thesis]. University of Adelaide; 2019. Available from: http://hdl.handle.net/2440/124703
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Princeton University
26.
Ranganath, Rajesh.
Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications
.
Degree: PhD, 2017, Princeton University
URL: http://arks.princeton.edu/ark:/88435/dsp01pr76f608w
► Probabilistic generative models are robust to noise, uncover unseen patterns, and make predictions about the future. These models have been used successfully to solve problems…
(more)
▼ Probabilistic generative models are robust to noise, uncover unseen patterns, and make predictions about the future. These models have been used successfully to solve problems in neuroscience, astrophysics, genetics, and medicine. The main computational challenge is computing the hidden structure given the data – posterior
inference. For most models of interest, computing the posterior distribution requires approximations like variational
inference. Variational
inference transforms posterior
inference into optimization. Classically, this optimization problem was feasible to deploy in only a small fraction of models.
This thesis develops black box variational
inference. Black box variational
inference is a variational
inference algorithm that is easy to deploy on a broad class of models and has already found use in models for neuroscience and health care. It makes new kinds of models possible, ones that were too unruly for previous
inference methods.
One set of models we develop is deep exponential families. Deep exponential families uncover new kinds of hidden pattens while being predictive of future data. Many existing models are deep exponential families. Black box variational
inference makes it possible for us to quickly study a broad range of deep exponential families with minimal added effort for each new type of deep exponential family.
The ideas around black box variational
inference also facilitate new kinds of variational methods. First, we develop hierarchical variational models. Hierarchical variational models improve the approximation quality of variational
inference by building higher-fidelity approximations from coarser ones. We show that they help with
inference in deep exponential families. Second, we introduce operator variational
inference. Operator variational
inference delves into the possible distance measures that can be used for the variational optimization problem. We show that this formulation categorizes various variational
inference methods and enables variational approximations without tractable densities.
By developing black box variational
inference, we have opened doors to new models, better posterior approximations, and new varieties of variational
inference algorithms.
Advisors/Committee Members: Blei, David M (advisor).
Subjects/Keywords: Bayesian Statistics;
Machine Learning;
Variational Inference
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Ranganath, R. (2017). Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications
. (Doctoral Dissertation). Princeton University. Retrieved from http://arks.princeton.edu/ark:/88435/dsp01pr76f608w
Chicago Manual of Style (16th Edition):
Ranganath, Rajesh. “Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications
.” 2017. Doctoral Dissertation, Princeton University. Accessed March 04, 2021.
http://arks.princeton.edu/ark:/88435/dsp01pr76f608w.
MLA Handbook (7th Edition):
Ranganath, Rajesh. “Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications
.” 2017. Web. 04 Mar 2021.
Vancouver:
Ranganath R. Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications
. [Internet] [Doctoral dissertation]. Princeton University; 2017. [cited 2021 Mar 04].
Available from: http://arks.princeton.edu/ark:/88435/dsp01pr76f608w.
Council of Science Editors:
Ranganath R. Black Box Variational Inference: Scalable, Generic Bayesian Computation and its Applications
. [Doctoral Dissertation]. Princeton University; 2017. Available from: http://arks.princeton.edu/ark:/88435/dsp01pr76f608w
Universitat de Valencia
27.
Martínez Minaya, Joaquín.
Recent statistical advances and applications of species distribution modeling
.
Degree: 2019, Universitat de Valencia
URL: http://hdl.handle.net/10550/71315
► En el mundo en que vivimos, producimos aproximadamente 2.5 quintillones de bytes de datos por día. Esta enorme cantidad de datos proviene de las redes…
(more)
▼ En el mundo en que vivimos, producimos aproximadamente 2.5 quintillones de bytes de datos por día. Esta enorme cantidad de datos proviene de las redes sociales, Internet, satélites, etc. Todos estos datos, que se pueden registrar en el tiempo o en el espacio, son información que puede ayudarnos a comprender la propagación de una enfermedad, el movimiento de especies o el cambio climático. El uso de modelos estadísticos complejos ha aumentado recientemente en el contexto del estudio de la distribución de especies. Esta complejidad ha hecho que los procesos inferenciales y predictivos sean difíciles de realizar. El enfoque bayesiano se ha convertido en una buena opción para lidiar con estos modelos, debido a la facilidad con la que se puede incorporar la información previa, junto con el hecho de que proporciona una estimación de la incertidumbre más realista y precisa.
En esta tesis, mostramos una visión actualizada del uso de las últimas herramientas estadísticas que han surgido en la aplicación de modelos de distribución de especies (SDMs) en contextos reales desde una perspectiva bayesiana, y desarrollamos nuevas herramientas metodológicas para resolver algunos problemas estadísticos que aparecieron en ese proceso.
Con respecto a la aplicación de las últimas herramientas estadísticas en el contexto de los SDMs, los objetivos específicos han sido modelizar la producción de ascosporas Plurivorosphaerella nawae en la hojarasca de caqui; estudiar los factores espaciales y climáticos asociados con la distribución de la mancha negra de los cítricos causada por el hongo Phyllosticta citricarpa; analizar los efectos de la estructura genética y la autocorrelación espacial en los cambios de rango de distribución de las especies; y estudiar la distribución del delfín mular (Tursiops truncatus). Dos objetivos han marcado la parte más metodológica de la tesis: una revisión centrada en los problemas estadísticos en SDMs y la implementación de la regresión de Dirichlet bayesiana en el contexto de la aproximación de Laplace anidada integrada (INLA).
La tesis que aquí presentamos es un compendio de ocho artículos y a continuación mostramos su estructura. En los cuatro primeros capítulos presentamos una introducción general que incluye una descripción de los objetivos (Capítulo 1), la base de la metodología empleada (Capítulos 2 y 3) y una descripción de los resultados obtenidos (Capítulo 4). En los ocho capítulos siguientes, mostramos todos los artículos que componen este compendio. Y por último, incluimos el Capítulo 13, donde se presentan algunas conclusiones y líneas futuras de investigación, seguido de una bibliografía genérica correspondiente a los capítulos introductorios.
Advisors/Committee Members: Conesa Guillén, David (advisor).
Subjects/Keywords: bayesian inference;
inla;
species distribution models;
geostatistics
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Martínez Minaya, J. (2019). Recent statistical advances and applications of species distribution modeling
. (Doctoral Dissertation). Universitat de Valencia. Retrieved from http://hdl.handle.net/10550/71315
Chicago Manual of Style (16th Edition):
Martínez Minaya, Joaquín. “Recent statistical advances and applications of species distribution modeling
.” 2019. Doctoral Dissertation, Universitat de Valencia. Accessed March 04, 2021.
http://hdl.handle.net/10550/71315.
MLA Handbook (7th Edition):
Martínez Minaya, Joaquín. “Recent statistical advances and applications of species distribution modeling
.” 2019. Web. 04 Mar 2021.
Vancouver:
Martínez Minaya J. Recent statistical advances and applications of species distribution modeling
. [Internet] [Doctoral dissertation]. Universitat de Valencia; 2019. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/10550/71315.
Council of Science Editors:
Martínez Minaya J. Recent statistical advances and applications of species distribution modeling
. [Doctoral Dissertation]. Universitat de Valencia; 2019. Available from: http://hdl.handle.net/10550/71315
University of Sydney
28.
Babbar, Sakshi.
Inferring Anomalies from Data using Bayesian Networks
.
Degree: 2013, University of Sydney
URL: http://hdl.handle.net/2123/9371
► Existing studies on data mining has largely focused on the design of measures and algorithms to identify outliers in large and high dimensional categorical and…
(more)
▼ Existing studies on data mining has largely focused on the design of measures and algorithms to identify outliers in large and high dimensional categorical and numeric databases. However, not much stress has been given on the interestingness of the reported outlier. One way to ascertain interestingness and usefulness of the reported outlier is by making use of domain knowledge. In this thesis, we present measures to discover outliers based on background knowledge, represented by a Bayesian network. Using causal relationships between attributes encoded in the Bayesian framework, we demonstrate that meaningful outliers, i.e., outliers which encode important or new information are those which violate causal relationships encoded in the model. Depending upon nature of data, several approaches are proposed to identify and explain anomalies using Bayesian knowledge. Outliers are often identified as data points which are ``rare'', ''isolated'', or ''far away from their nearest neighbors''. We show that these characteristics may not be an accurate way of describing interesting outliers. Through a critical analysis on several existing outlier detection techniques, we show why there is a mismatch between outliers as entities described by these characteristics and ``real'' outliers as identified using Bayesian approach. We show that the Bayesian approaches presented in this thesis has better accuracy in mining genuine outliers while, keeping a low false positive rate as compared to traditional outlier detection techniques.
Subjects/Keywords: Bayesian networks;
outlier;
causal inference;
knowledge discovery
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Babbar, S. (2013). Inferring Anomalies from Data using Bayesian Networks
. (Thesis). University of Sydney. Retrieved from http://hdl.handle.net/2123/9371
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Babbar, Sakshi. “Inferring Anomalies from Data using Bayesian Networks
.” 2013. Thesis, University of Sydney. Accessed March 04, 2021.
http://hdl.handle.net/2123/9371.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Babbar, Sakshi. “Inferring Anomalies from Data using Bayesian Networks
.” 2013. Web. 04 Mar 2021.
Vancouver:
Babbar S. Inferring Anomalies from Data using Bayesian Networks
. [Internet] [Thesis]. University of Sydney; 2013. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2123/9371.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Babbar S. Inferring Anomalies from Data using Bayesian Networks
. [Thesis]. University of Sydney; 2013. Available from: http://hdl.handle.net/2123/9371
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
University of Sydney
29.
Tompkins, Anthony.
Bayesian Spatio-Temporal Modelling with Fourier Features
.
Degree: 2018, University of Sydney
URL: http://hdl.handle.net/2123/21328
► One of the most powerful machine learning techniques is \emph{Gaussian Processes} (GPs) which incur an O(N3) complexity in the number of data samples. In regression…
(more)
▼ One of the most powerful machine learning techniques is \emph{Gaussian Processes} (GPs) which incur an O(N3) complexity in the number of data samples. In regression and classification there exist approximation methods which typically rely on M \emph{inducing points} but still typically incur an O(NM2) complexity in the data and corresponding inducing points which have reduced expressiveness the larger the dataset becomes. These methods are typically unable to learn if the number of datapoints becomes computationally intractable. It is this limitation of traditional methods that invites us to explore alternative representations of kernels to enable scalable modelling and inference for spatio-temporal phenomena. The key insight we leverage is providing \emph{feature}-space representations of kernels which have computational dependence \emph{independent} of data samples. While such representations exist in various forms, they typically address kernels of infinite support and have not been investigated extensively for modeling periodicity or data supported on bounded intervals. Our approach leverages methods in harmonic analysis to provide an alternative form of representing kernels using Fourier Series which we demonstrate to have superior performance to alternative feature representations. Our methodology further develops \emph{compositional} kernels and show it is straightforward to integrate our Fourier series features with standard kernels. With compositions of kernels we are able to represent nuances in the data that canonical kernels typically cannot represent. This thesis brings the following contributions: 1) A new formulation of representing univariate periodic kernels using Fourier series that allows one to perform scalable inference with a large number of samples; 2) A generalisation of univariate periodic kernels into the multivariate domain which allows tractable higher dimensional inference; 3) An efficient method for the tricky problem of seeding and learning periodic hyperparameters; 4) A generalised framework that allows one to perform compositional kernel learning in a Bayesian framework for spatio-temporal phenomena.
Subjects/Keywords: Gaussian Processes;
bayesian inference;
periodicity;
Timeseries
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Tompkins, A. (2018). Bayesian Spatio-Temporal Modelling with Fourier Features
. (Thesis). University of Sydney. Retrieved from http://hdl.handle.net/2123/21328
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Chicago Manual of Style (16th Edition):
Tompkins, Anthony. “Bayesian Spatio-Temporal Modelling with Fourier Features
.” 2018. Thesis, University of Sydney. Accessed March 04, 2021.
http://hdl.handle.net/2123/21328.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
MLA Handbook (7th Edition):
Tompkins, Anthony. “Bayesian Spatio-Temporal Modelling with Fourier Features
.” 2018. Web. 04 Mar 2021.
Vancouver:
Tompkins A. Bayesian Spatio-Temporal Modelling with Fourier Features
. [Internet] [Thesis]. University of Sydney; 2018. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2123/21328.
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
Council of Science Editors:
Tompkins A. Bayesian Spatio-Temporal Modelling with Fourier Features
. [Thesis]. University of Sydney; 2018. Available from: http://hdl.handle.net/2123/21328
Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation
University of Texas – Austin
30.
Zhang, Michael Minyi.
Scalable inference for Bayesian non-parametrics.
Degree: PhD, Statistics, 2018, University of Texas – Austin
URL: http://hdl.handle.net/2152/65734
► Bayesian non-parametric models, despite their theoretical elegance, face a serious computational burden that prevents their use in serious "big data'' scenarios. Furthermore, we cannot expect…
(more)
▼ Bayesian non-parametric models, despite their theoretical elegance, face a serious computational burden that prevents their use in serious "big data'' scenarios. Furthermore, we cannot expect the data in "big data'' to exist solely on one processor, so we must have parallel algorithms that are valid
Bayesian inference samplers. However, inherent dependencies in
Bayesian non-parametric models make this task very difficult. Instead, we must either construct good approximations or develop clever reformulations of our models so that we perform
inference with provably accurate results. This thesis will discuss four methods developed to parallelize
inference in the
Bayesian and
Bayesian non-parametric setting.
Advisors/Committee Members: Williamson, Sinead (advisor), Mueller, Peter (committee member), Scott, James G (committee member), Xing, Eric P (committee member).
Subjects/Keywords: Bayesian non-parametrics; Scalable inference; Machine learning
Record Details
Similar Records
Cite
Share »
Record Details
Similar Records
Cite
« Share
❌
APA ·
Chicago ·
MLA ·
Vancouver ·
CSE |
Export
to Zotero / EndNote / Reference
Manager
APA (6th Edition):
Zhang, M. M. (2018). Scalable inference for Bayesian non-parametrics. (Doctoral Dissertation). University of Texas – Austin. Retrieved from http://hdl.handle.net/2152/65734
Chicago Manual of Style (16th Edition):
Zhang, Michael Minyi. “Scalable inference for Bayesian non-parametrics.” 2018. Doctoral Dissertation, University of Texas – Austin. Accessed March 04, 2021.
http://hdl.handle.net/2152/65734.
MLA Handbook (7th Edition):
Zhang, Michael Minyi. “Scalable inference for Bayesian non-parametrics.” 2018. Web. 04 Mar 2021.
Vancouver:
Zhang MM. Scalable inference for Bayesian non-parametrics. [Internet] [Doctoral dissertation]. University of Texas – Austin; 2018. [cited 2021 Mar 04].
Available from: http://hdl.handle.net/2152/65734.
Council of Science Editors:
Zhang MM. Scalable inference for Bayesian non-parametrics. [Doctoral Dissertation]. University of Texas – Austin; 2018. Available from: http://hdl.handle.net/2152/65734
◁ [1] [2] [3] [4] [5] … [28] ▶
. 
Open Access Theses and Dissertations
