Advanced search options

Advanced Search Options 🞨

Browse by author name (“Author name starts with…”).

Find ETDs with:

in
/  
in
/  
in
/  
in

Written in Published in Earliest date Latest date

Sorted by

Results per page:

Sorted by: relevance · author · university · dateNew search

You searched for +publisher:"Universidade Federal do Amazonas" +contributor:("Carvalho, Andr? Luiz da Costa"). Showing records 1 – 2 of 2 total matches.

Search Limiters

Last 2 Years | English Only

No search limiters apply to these results.

▼ Search Limiters

1. Silveira, Denys Dion?sio Bezerra. Modelos de T?picos baseados em Autocodificadores Variacionais utilizando as distribui??es Gumbel-Softmax e mistura de Normais-Log?sticas.

Degree: 2018, Universidade Federal do Amazonas

Modelos probabil?sticos de t?picos s?o modelos estat?sticos capazes de identificar t?picos em uma cole??o de texto. Eles s?o amplamente aplicados em tarefas relacionadas ? ?rea de Processamento de Linguagem Natural, uma vez que capturam com sucesso rela??es latentes por meio da an?lise de dados n?o rotulados. Entretanto, solu??es anal?ticas para a infer?ncia Bayesiana desses modelos s?o geralmente intrat?veis, dificultando a proposta de modelos probabil?sticos que sejam mais expressivos. Neste cen?rio, os Autocodificadores Variacionais (ACVs), m?todos que empregam uma rede de infer?ncia baseada em redes neurais respons?vel por estimar a distribui??o a posteriori, tornaram-se uma alternativa promissora para inferir distribui??es de t?picos em cole??es de texto. Estes modelos, contudo, tamb?m introduzem novos desafios, tal como a necessidade de distribui??es cont?nuas e reparametriz?veis que podem n?o se ajustar ?s distribui??es reais dos t?picos. Al?m disso, redes de infer?ncia tendem a apresentar um problema conhecido como colapso de componentes, onde apenas alguns t?picos contendo poucos termos correlacionados s?o efetivamente extra?dos. Para tentar evitar estes problemas, prop?em-se dois novos m?todos de t?picos. O primeiro (GSDTM) ? baseado em uma distribui??o cont?nua pseudocateg?rica denominada Gumbel-Softmax, capaz de gerar amostras aproximadamente categ?ricas, enquanto o segundo (LMDTM) adota uma mistura de distribui??es Normais-log?sticas, que pode ser adequada em cen?rios onde a distribui??o dos dados ? complexa. Apresenta-se tamb?m um estudo sobre o impacto que diferentes escolhas de modelagem t?m sobre os t?picos gerados, observando um compromisso entre coer?ncia dos t?picos e a qualidade do modelo gerador. Por meio de experimentos usando duas cole??es de dados de refer?ncia, tr?s m?tricas distintas de avalia??o quantitativa e uma inspe??o qualitativa, mostra-se que o modelo GSDTM supera de forma significativa os modelos de t?picos considerados estado da arte em grande parte dos cen?rios de teste, em termos de coer?ncia m?dia de t?picos e perplexidade.

Probabilistic topic models are statistical models which are able to identify topics on textual data. They are widely applied in many tasks related to Natural Language Processing due to their effective use of unlabeled data to capture latent relations. Analytical solutions for Bayesian inference of such models, however, are usually intractable, hindering the proposition of highly expressive text models. In this scenario, Variational Auto-Encoders (VAEs), where an artificial neural-based inference network is used to approximate the posterior distribution, became a promising alternative for inferring latent topic distributions of text documents. These models, however, also pose new challenges such as the requirement of continuous and reparameterizable distributions which may not fit so well the true latent topic distributions. Moreover, inference networks are prone to a well-known problem called component collapsing, where a little number of topics are…

Advisors/Committee Members: Cristo, Marco Ant?nio Pinheiro de, [email protected], http://lattes.cnpq.br/6261175351521953, Carvalho, Andr? Luiz da Costa, http://lattes.cnpq.br/4863447798119856, Colonna, Juan Gabriel, http://lattes.cnpq.br/9535853909210803, Pappa, Gisele Lobo, http://lattes.cnpq.br/5936682335701497, Carvalho, Andr? Luiz da Costa, http://lattes.cnpq.br/4863447798119856, [email protected].

Subjects/Keywords: Redes neurais (Computa??o); Teoria bayesiana de decis?o estat?stica; CI?NCIAS EXATAS E DA TERRA; CI?NCIAS EXATAS E DA TERRA: CI?NCIA DA COMPUTA??O; Modelos de T?picos; Autocodificadores Variacionais; Infer?ncia Bayesiana; Aprendizagem Profunda

Record DetailsSimilar RecordsGoogle PlusoneFacebookTwitterCiteULikeMendeleyreddit

APA · Chicago · MLA · Vancouver · CSE | Export to Zotero / EndNote / Reference Manager

APA (6th Edition):

Silveira, D. D. B. (2018). Modelos de T?picos baseados em Autocodificadores Variacionais utilizando as distribui??es Gumbel-Softmax e mistura de Normais-Log?sticas. (Masters Thesis). Universidade Federal do Amazonas. Retrieved from https://tede.ufam.edu.br/handle/tede/7439

Chicago Manual of Style (16th Edition):

Silveira, Denys Dion?sio Bezerra. “Modelos de T?picos baseados em Autocodificadores Variacionais utilizando as distribui??es Gumbel-Softmax e mistura de Normais-Log?sticas.” 2018. Masters Thesis, Universidade Federal do Amazonas. Accessed August 11, 2020. https://tede.ufam.edu.br/handle/tede/7439.

MLA Handbook (7th Edition):

Silveira, Denys Dion?sio Bezerra. “Modelos de T?picos baseados em Autocodificadores Variacionais utilizando as distribui??es Gumbel-Softmax e mistura de Normais-Log?sticas.” 2018. Web. 11 Aug 2020.

Vancouver:

Silveira DDB. Modelos de T?picos baseados em Autocodificadores Variacionais utilizando as distribui??es Gumbel-Softmax e mistura de Normais-Log?sticas. [Internet] [Masters thesis]. Universidade Federal do Amazonas; 2018. [cited 2020 Aug 11]. Available from: https://tede.ufam.edu.br/handle/tede/7439.

Council of Science Editors:

Silveira DDB. Modelos de T?picos baseados em Autocodificadores Variacionais utilizando as distribui??es Gumbel-Softmax e mistura de Normais-Log?sticas. [Masters Thesis]. Universidade Federal do Amazonas; 2018. Available from: https://tede.ufam.edu.br/handle/tede/7439

2. Fonseca, Paulo C?sar da Rocha; http://lattes.cnpq.br/3639575844521754. A deep learning framework for BGP anomaly detection and classification.

Degree: 2019, Universidade Federal do Amazonas

The Border Gateway Protocol (BGP) is the default Internet routing protocol that manages connectivity among Autonomous Systems (ASes). Although BGP disruptions are rare when they occur the consequences can be very damaging. Consequently, there has been a considerable effort aimed at understanding what is normal and abnormal BGP traffic and, in so doing, enable potentially disruptive anomalous traffic to be identified quickly. Even though there is an extensive research on anomaly detection, there are two major gaps in current literature: the scarcity of public datasets for all types of events and the lack of a BGP anomaly classification framework that differentiates anomaly classes. Since that there are no public datasets of labeled BGP anomalous events, each model was validated using different datasets, which had to be individually generated for each approach. The absence of common groundwork dataset increases the difficulty in comparing different approaches. The lack of a classification framework hinders the deployment of specific mitigation measures to each anomaly class in an automated fashion. In the current work, we address both problems: 1) We provide a BGP dataset generation tool and publicly available datasets for different anomaly classes. These datasets contain the most used features by previous research efforts and additional novel features; 2) We address the BGP anomaly classification problem by developing a framework that uses deep learning as the core engine of an anomaly detection and classification mechanism. We built a model that exploits different neural network architectures advantages. Both novel features and the BGP anomaly detector and classifier were evaluated and it was demonstrated that they can be used to react to anomalies in real-time and leverage the deployment of different mitigation and coordination strategies to different anomaly classes in an autonomous fashion.

The Border Gateway Protocol (BGP) is the default Internet routing protocol that manages connectivity among Autonomous Systems (ASes). Although BGP disruptions are rare when they occur the consequences can be very damaging. Consequently, there has been a considerable effort aimed at understanding what is normal and abnormal BGP traffic and, in so doing, enable potentially disruptive anomalous traffic to be identified quickly. Even though there is an extensive research on anomaly detection, there are two major gaps in current literature: the scarcity of public datasets for all types of events and the lack of a BGP anomaly classification framework that differentiates anomaly classes. Since that there are no public datasets of labeled BGP anomalous events, each model was validated using different datasets, which had to be individually generated for each approach. The absence of common groundwork dataset increases the difficulty in comparing different approaches. The lack of a classification framework hinders the deployment of specific mitigation measures to each anomaly class in an automated fashion. In the current work, we…

Advisors/Committee Members: Mota, Edjard Souza, http://lattes.cnpq.br/0757666181169076, Feitosa, Eduardo Luzeiro, http://lattes.cnpq.br/5939944067207881, Carvalho, Andr? Luiz da Costa, http://lattes.cnpq.br/4863447798119856, Souza, Jose Neuman de, http://lattes.cnpq.br/3614256141054800.

Subjects/Keywords: Border Gateway Protocol; Machine Learning; Dataset generation; Autonomous Systems; Anomalias BGP; CI?NCIAS EXATAS E DA TERRA: CI?NCIA DA COMPUTA??O; Border Gateway Protocol; Anomaly detection; Machine Learning; Dataset generation; Detec??o de anomalias

Record DetailsSimilar RecordsGoogle PlusoneFacebookTwitterCiteULikeMendeleyreddit

APA · Chicago · MLA · Vancouver · CSE | Export to Zotero / EndNote / Reference Manager

APA (6th Edition):

Fonseca, P. C. d. R. h. c. b. (2019). A deep learning framework for BGP anomaly detection and classification. (Doctoral Dissertation). Universidade Federal do Amazonas. Retrieved from https://tede.ufam.edu.br/handle/tede/7700

Chicago Manual of Style (16th Edition):

Fonseca, Paulo C?sar da Rocha; http://lattes cnpq br/3639575844521754. “A deep learning framework for BGP anomaly detection and classification.” 2019. Doctoral Dissertation, Universidade Federal do Amazonas. Accessed August 11, 2020. https://tede.ufam.edu.br/handle/tede/7700.

MLA Handbook (7th Edition):

Fonseca, Paulo C?sar da Rocha; http://lattes cnpq br/3639575844521754. “A deep learning framework for BGP anomaly detection and classification.” 2019. Web. 11 Aug 2020.

Vancouver:

Fonseca PCdRhcb. A deep learning framework for BGP anomaly detection and classification. [Internet] [Doctoral dissertation]. Universidade Federal do Amazonas; 2019. [cited 2020 Aug 11]. Available from: https://tede.ufam.edu.br/handle/tede/7700.

Council of Science Editors:

Fonseca PCdRhcb. A deep learning framework for BGP anomaly detection and classification. [Doctoral Dissertation]. Universidade Federal do Amazonas; 2019. Available from: https://tede.ufam.edu.br/handle/tede/7700

.