subject:(Explore Then Commit Bandits)

University of Illinois – Urbana-Champaign

1. Yekkehkhany, Ali. Risk-averse multi-armed bandits and game theory.

Degree: PhD, Electrical & Computer Engr, 2020, University of Illinois – Urbana-Champaign

► The multi-armed bandit (MAB) and game theory literature is mainly focused on the expected cumulative reward and the expected payoffs in a game, respectively. In… (more)

▼ The multi-armed bandit (MAB) and game theory literature is mainly focused on the expected cumulative reward and the expected payoffs in a game, respectively. In contrast, the rewards and the payoffs are often random variables whose expected values only capture a vague idea of the overall distribution. The focus of this dissertation is to study the fundamental limits of the existing bandits and game theory problems in a risk-averse framework and propose new ideas that address the shortcomings. The author believes that human beings are mostly risk-averse, so studying multi-armed bandits and game theory from the point of view of risk aversion, rather than expected reward/payoff, better captures reality. In this manner, a specific class of multi-armed bandits, called explore-then-commit bandits, and stochastic games are studied in this dissertation, which are based on the notion of Risk-Averse Best Action Decision with Incomplete Information (R-ABADI, Abadi is the maiden name of the author's mother). The goal of the classical multi-armed bandits is to exploit the arm with the maximum score defined as the expected value of the arm reward. Instead, we propose a new definition of score that is derived from the joint distribution of all arm rewards and captures the reward of an arm relative to those of all other arms. We use a similar idea for games and propose a risk-averse R-ABADI equilibrium in game theory that is possibly different from the Nash equilibrium. The payoff distributions are taken into account to derive the risk-averse equilibrium, while the expected payoffs are used to find the Nash equilibrium. The fundamental properties of games, e.g. pure and mixed risk-averse R-ABADI equilibrium and strict dominance, are studied in the new framework and the results are expanded to finite-time games. Furthermore, the stochastic congestion games are studied from a risk-averse perspective and three classes of equilibria are proposed for such games. It is shown by examples that the risk-averse behavior of travelers in a stochastic congestion game can improve the price of anarchy in Pigou and Braess networks. Furthermore, the Braess paradox does not occur to the extent proposed originally when travelers are risk-averse. We also study an online affinity scheduling problem with no prior knowledge of the task arrival rates and processing rates of different task types on different servers. We propose the Blind GB-PANDAS algorithm that utilizes an exploration-exploitation scheme to load balance incoming tasks on servers in an online fashion. We prove that Blind GB-PANDAS is throughput optimal, i.e. it stabilizes the system as long as the task arrival rates are inside the capacity region. The Blind GB-PANDAS algorithm is compared to FCFS, Max-Weight, and c-mu-rule algorithms in terms of average task completion time through simulations, where the same exploration-exploitation approach as Blind GB-PANDAS is used for Max-Weight and c-μ-rule. The extensive simulations show that the Blind GB-PANDAS algorithm conspicuously… Advisors/Committee Members: Nagi, Rakesh (advisor), Nagi, Rakesh (Committee Chair), Hajek, Bruce (committee member), Shomorony, Ilan (committee member), Srikant, Rayadurgam (committee member).

Subjects/Keywords: Online Learning; Multi-Armed Bandits; Exploration-Exploitation; Explore-Then-Commit Bandits; Risk-Aversion; Game Theory; Stochastic Game Theory; Congestion Games; Affinity Scheduling; MapReduce; Data Center

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Yekkehkhany, A. (2020). Risk-averse multi-armed bandits and game theory. (Doctoral Dissertation). University of Illinois – Urbana-Champaign. Retrieved from http://hdl.handle.net/2142/108439

Chicago Manual of Style (16^th Edition):

Yekkehkhany, Ali. “Risk-averse multi-armed bandits and game theory.” 2020. Doctoral Dissertation, University of Illinois – Urbana-Champaign. Accessed April 10, 2021. http://hdl.handle.net/2142/108439.

MLA Handbook (7^th Edition):

Yekkehkhany, Ali. “Risk-averse multi-armed bandits and game theory.” 2020. Web. 10 Apr 2021.

Vancouver:

Yekkehkhany A. Risk-averse multi-armed bandits and game theory. [Internet] [Doctoral dissertation]. University of Illinois – Urbana-Champaign; 2020. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/2142/108439.

Council of Science Editors:

Yekkehkhany A. Risk-averse multi-armed bandits and game theory. [Doctoral Dissertation]. University of Illinois – Urbana-Champaign; 2020. Available from: http://hdl.handle.net/2142/108439

2. Reis, Valentin. Apprentissage pour le contrôle de plateformes parallèles à large échelle : Learning to control large-scale parallel platforms.

Degree: Docteur es, Informatique, 2018, Université Grenoble Alpes (ComUE)

URL: http://www.theses.fr/2018GREAM045

►

Fournir les infrastructures de calcul nécessaires à la résolution des problèmescom-plexes de la société moderne constitue un défistratégique. Lesorganisations y répondent classiquement en mettant en… (more)

▼

Fournir les infrastructures de calcul nécessaires à la résolution des problèmescom-plexes de la société moderne constitue un défistratégique. Lesorganisations y répondent classiquement en mettant en place de largesinfrastructures de calcul parallèle et distribué. Les vendeurs de systèmes deCalcul Hautes Performances sont incités par la compétition à produire toujoursplus de puissance de calcul et de stockage, ce qui mène à des plateformes”Petascale“ spécifiques et sophistiquées, et bientôt à des machines”Exascale“. Ces systèmes sont gérés de manière centralisée à l’aide desolutions logicielles de gestion de jobs et de resources dédiées. Un problèmecrucial auquel répondent ces logiciels est le problème d’ordonnancement, pourlequel le gestionnaire de resources doit choisir quand, et sur quellesresources exécuter quelle tache calculatoire. Cette thèse fournit des solutionsà ce problème. Toutes les plateformes sont différentes. En effet, leurinfrastructure, le comportement de leurs utilisateurs et les objectifs del’organisation hôte varient. Nous soutenons donc que les politiquesd’ordonnancement doivent s’adapter au comportement des systèmes. Dans cemanuscrit, nous présentons plusieurs manières d’obtenir cette adaptativité. Atravers une approche expérimentale, nous étudions plusieurs compromis entre lacomplexité de l’approche, le gain potentiel, et les risques pris.

Providing the computational infrastucture needed to solve complex problemsarising in modern society is a strategic challenge. Organisations usuallyadress this problem by building extreme-scale parallel and distributedplatforms. High Performance Computing (HPC) vendors race for more computingpower and storage capacity, leading to sophisticated specific Petascaleplatforms, soon to be Exascale platforms. These systems are centrally managedusing dedicated software solutions called Resource and Job Management Systems(RJMS). A crucial problem adressed by this software layer is the job schedulingproblem, where the RJMS chooses when and on which resources computational taskswill be executed. This manuscript provides ways to adress this schedulingproblem. No two platforms are identical. Indeed, the infrastructure, userbehavior and organization's goals all change from one system to the other. Wetherefore argue that scheduling policies should be adaptative to the system'sbehavior. In this manuscript, we provide multiple ways to achieve thisadaptativity. Through an experimental approach, we study various tradeoffsbetween the complexity of the approach, the potential gain, and the riskstaken.

Advisors/Committee Members: Trystram, Denis (thesis director), Lelong, Jérôme (thesis director).

Subjects/Keywords: Allocation de resources; Bandits; Ordonnancement; Bandits; Scheduling; Bandits; 004

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Reis, V. (2018). Apprentissage pour le contrôle de plateformes parallèles à large échelle : Learning to control large-scale parallel platforms. (Doctoral Dissertation). Université Grenoble Alpes (ComUE). Retrieved from http://www.theses.fr/2018GREAM045

Chicago Manual of Style (16^th Edition):

Reis, Valentin. “Apprentissage pour le contrôle de plateformes parallèles à large échelle : Learning to control large-scale parallel platforms.” 2018. Doctoral Dissertation, Université Grenoble Alpes (ComUE). Accessed April 10, 2021. http://www.theses.fr/2018GREAM045.

MLA Handbook (7^th Edition):

Reis, Valentin. “Apprentissage pour le contrôle de plateformes parallèles à large échelle : Learning to control large-scale parallel platforms.” 2018. Web. 10 Apr 2021.

Vancouver:

Reis V. Apprentissage pour le contrôle de plateformes parallèles à large échelle : Learning to control large-scale parallel platforms. [Internet] [Doctoral dissertation]. Université Grenoble Alpes (ComUE); 2018. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2018GREAM045.

Council of Science Editors:

Reis V. Apprentissage pour le contrôle de plateformes parallèles à large échelle : Learning to control large-scale parallel platforms. [Doctoral Dissertation]. Université Grenoble Alpes (ComUE); 2018. Available from: http://www.theses.fr/2018GREAM045

Addis Ababa University

3. Mulu, Woldemariam. Gender Mainstreaming Aspects in the Women’s Affairs and Economic and Business Affairs General Directorates of Ministry of Foreign Affairs, Addis Ababa .

Degree: 2009, Addis Ababa University

URL: http://etd.aau.edu.et/dspace/handle/123456789/6706

► The study is designed to explore the status of gender mainstreaming aspects in two general directorates of Ministry of Foreign Affairs (MOFA): Women’s Affairs General… (more)

▼ The study is designed to explore the status of gender mainstreaming aspects in two general directorates of Ministry of Foreign Affairs (MOFA): Women’s Affairs General Directorates (WAGD) and Economic Business Affairs General Directorates (EBAGD). Particularly, the study targeted on three departments of WAGD and two departments of EBAGD: The departments of WAGD are Resource Mobilization, Plan Research and Capacity Building Department (RMRPCBD), Policy and Affirmative Action Affairs Department (PAAAD), and Women’s and Children’s Trafficking Department. The departments of EBAGD are Foreign Trade, Investments and Tourism Development (FTITDD) and Mobilization of Resources for Development and Technical Cooperation Research and plan Department (MRDTCRPD). The study used qualitative approach where it applied in-depth interview, focus group discussion, and personal observation of the researcher to collect the needed information. Furthermore, document analysis was used as part of the study, to find out gender dimensions and gaps in MOFA. Based on the data analysis, the study findings revealed the following results. With regard to policy and affirmative action, the results indicated that at present there is no gender policy, strategy, or guideline that guides to mainstream gender in MOFA. However, at the moment, WAGD of MOFA is responsible to formulate and it is on the process of developing gender policy or guideline. In the foreign policy and strategy document, gender aspects are not integrated. Hence, it is not gender sensitive. On the other hand, MOFA’s commitment is reflected in that it has developed gender affirmative implementation guide line concerned with women employees but not for women external customers. In spite of the formulation of affirmative guideline, the findings showed that little was done to increase the number of women employees in MOFA. The result also indicated MOFA allocates budgets to WAGD for gender mainstreaming purposes at the organizational level. With regard to protecting women from trafficking, the result showed the management of MOFA has established a department called “Women’s and Children’s Trafficking Department“ with specific duties and responsibilities. Based on this, the department has started performing related activities. With respect to gender roles, the result reflected that higher managerial positions are held by men. As the data shows, the first four higher position in the head office and the embassies are occupied by women. Whereas the lower position such as secretary and cleaner are occupied by women. So in general, gender roles are stereotyped. Although there are some measures taken to translate the affirmative action given in the Ethiopian constitution and gender mainstreaming conventions to promote women in foreign trade investment and tourism, the affirmative action opportunities are not fully translated into action, according to the study. Furthermore, in the foreign resource mobilization (foreign loans, aids or technical support), no percentage… Advisors/Committee Members: Dr. Vijaya Subramaniyam (advisor).

Subjects/Keywords: designed to explore the status

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Mulu, W. (2009). Gender Mainstreaming Aspects in the Women’s Affairs and Economic and Business Affairs General Directorates of Ministry of Foreign Affairs, Addis Ababa . (Thesis). Addis Ababa University. Retrieved from http://etd.aau.edu.et/dspace/handle/123456789/6706

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Mulu, Woldemariam. “Gender Mainstreaming Aspects in the Women’s Affairs and Economic and Business Affairs General Directorates of Ministry of Foreign Affairs, Addis Ababa .” 2009. Thesis, Addis Ababa University. Accessed April 10, 2021. http://etd.aau.edu.et/dspace/handle/123456789/6706.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Mulu, Woldemariam. “Gender Mainstreaming Aspects in the Women’s Affairs and Economic and Business Affairs General Directorates of Ministry of Foreign Affairs, Addis Ababa .” 2009. Web. 10 Apr 2021.

Vancouver:

Mulu W. Gender Mainstreaming Aspects in the Women’s Affairs and Economic and Business Affairs General Directorates of Ministry of Foreign Affairs, Addis Ababa . [Internet] [Thesis]. Addis Ababa University; 2009. [cited 2021 Apr 10]. Available from: http://etd.aau.edu.et/dspace/handle/123456789/6706.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Mulu W. Gender Mainstreaming Aspects in the Women’s Affairs and Economic and Business Affairs General Directorates of Ministry of Foreign Affairs, Addis Ababa . [Thesis]. Addis Ababa University; 2009. Available from: http://etd.aau.edu.et/dspace/handle/123456789/6706

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Brno University of Technology

4. Piskořová, Zuzana. Traktografie axonálních svazků založená na multi-tensorových modelech: Multi-tensor model based tractography of axonal bundles.

Degree: 2019, Brno University of Technology

URL: http://hdl.handle.net/11012/65442

► The aim of this work is to design the tractography algorithm which consider microstructure features of the neuronal tissue. The methodological background is described, where… (more)

Subjects/Keywords: dMRI; odhad profilu anizotropni difuze; traktografie; COMMIT; optimalizace; dMRI; diffusion profile modeling; tractography; COMMIT; optimization

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Piskořová, Z. (2019). Traktografie axonálních svazků založená na multi-tensorových modelech: Multi-tensor model based tractography of axonal bundles. (Thesis). Brno University of Technology. Retrieved from http://hdl.handle.net/11012/65442

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Piskořová, Zuzana. “Traktografie axonálních svazků založená na multi-tensorových modelech: Multi-tensor model based tractography of axonal bundles.” 2019. Thesis, Brno University of Technology. Accessed April 10, 2021. http://hdl.handle.net/11012/65442.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Piskořová, Zuzana. “Traktografie axonálních svazků založená na multi-tensorových modelech: Multi-tensor model based tractography of axonal bundles.” 2019. Web. 10 Apr 2021.

Vancouver:

Piskořová Z. Traktografie axonálních svazků založená na multi-tensorových modelech: Multi-tensor model based tractography of axonal bundles. [Internet] [Thesis]. Brno University of Technology; 2019. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/11012/65442.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Piskořová Z. Traktografie axonálních svazků založená na multi-tensorových modelech: Multi-tensor model based tractography of axonal bundles. [Thesis]. Brno University of Technology; 2019. Available from: http://hdl.handle.net/11012/65442

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

University of Akron

5. Michaud, Heather M. Detection of Named Branch Origin for Git Commits.

Degree: MS, Computer Science, 2015, University of Akron

URL: http://rave.ohiolink.edu/etdc/view?acc_num=akron1436528915

► The named branch on which a change is committed in a Git repository providesvaluable insight into the evolution of a software project, including a natural… (more)

Subjects/Keywords: Computer Science; Git; repository mining; branching; merging; named branch; commit origin; software evolution; commit

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Michaud, H. M. (2015). Detection of Named Branch Origin for Git Commits. (Masters Thesis). University of Akron. Retrieved from http://rave.ohiolink.edu/etdc/view?acc_num=akron1436528915

Chicago Manual of Style (16^th Edition):

Michaud, Heather M. “Detection of Named Branch Origin for Git Commits.” 2015. Masters Thesis, University of Akron. Accessed April 10, 2021. http://rave.ohiolink.edu/etdc/view?acc_num=akron1436528915.

MLA Handbook (7^th Edition):

Michaud, Heather M. “Detection of Named Branch Origin for Git Commits.” 2015. Web. 10 Apr 2021.

Vancouver:

Michaud HM. Detection of Named Branch Origin for Git Commits. [Internet] [Masters thesis]. University of Akron; 2015. [cited 2021 Apr 10]. Available from: http://rave.ohiolink.edu/etdc/view?acc_num=akron1436528915.

Council of Science Editors:

Michaud HM. Detection of Named Branch Origin for Git Commits. [Masters Thesis]. University of Akron; 2015. Available from: http://rave.ohiolink.edu/etdc/view?acc_num=akron1436528915

University of Alberta

6. Wu, Yifan. Online Learning under Partial Feedback.

Degree: MS, Department of Computing Science, 2016, University of Alberta

URL: https://era.library.ualberta.ca/files/cvd66w003h

► In an online learning problem a player makes decisions in a sequential manner. In each round, the player receives some reward that depends on his… (more)

Subjects/Keywords: online learning; partial feedback; bandits

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Wu, Y. (2016). Online Learning under Partial Feedback. (Masters Thesis). University of Alberta. Retrieved from https://era.library.ualberta.ca/files/cvd66w003h

Chicago Manual of Style (16^th Edition):

Wu, Yifan. “Online Learning under Partial Feedback.” 2016. Masters Thesis, University of Alberta. Accessed April 10, 2021. https://era.library.ualberta.ca/files/cvd66w003h.

MLA Handbook (7^th Edition):

Wu, Yifan. “Online Learning under Partial Feedback.” 2016. Web. 10 Apr 2021.

Vancouver:

Wu Y. Online Learning under Partial Feedback. [Internet] [Masters thesis]. University of Alberta; 2016. [cited 2021 Apr 10]. Available from: https://era.library.ualberta.ca/files/cvd66w003h.

Council of Science Editors:

Wu Y. Online Learning under Partial Feedback. [Masters Thesis]. University of Alberta; 2016. Available from: https://era.library.ualberta.ca/files/cvd66w003h

7. Nicol, Olivier. Data-driven evaluation of contextual bandit algorithms and applications to dynamic recommendation : Évaluation basée sur des données d'algorithmes de bandits contextuels et application à la recommandation dynamique.

Degree: Docteur es, Informatique, 2014, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2014LIL10211

►

Ce travail de thèse a été réalisé dans le contexte de la recommandation dynamique. La recommandation est l'action de fournir du contenu personnalisé à un… (more)

▼

Ce travail de thèse a été réalisé dans le contexte de la recommandation dynamique. La recommandation est l'action de fournir du contenu personnalisé à un utilisateur utilisant une application, dans le but d'améliorer son utilisation e.g. la recommandation d'un produit sur un site marchant ou d'un article sur un blog. La recommandation est considérée comme dynamique lorsque le contenu à recommander ou encore les goûts des utilisateurs évoluent rapidement e.g. la recommandation d'actualités. Beaucoup d'applications auxquelles nous nous intéressons génèrent d'énormes quantités de données grâce à leurs millions d'utilisateurs sur Internet. Néanmoins, l'utilisation de ces données pour évaluer une nouvelle technique de recommandation ou encore comparer deux algorithmes de recommandation est loin d'être triviale. C'est cette problématique que nous considérons ici. Certaines approches ont déjà été proposées. Néanmoins elles sont très peu étudiées autant théoriquement (biais non quantifié, borne de convergence assez large...) qu'empiriquement (expériences sur données privées). Dans ce travail nous commençons par combler de nombreuses lacunes de l'analyse théorique. Ensuite nous discutons les résultats très surprenants d'une expérience à très grande échelle : une compétition ouverte au public que nous avons organisée. Cette compétition nous a permis de mettre en évidence une source de biais considérable et constamment présente en pratique : l'accélération temporelle. La suite de ce travail s'attaque à ce problème. Nous montrons qu'une approche à base de bootstrap permet de réduire mais surtout de contrôler ce biais.

The context of this thesis work is dynamic recommendation. Recommendation is the action, for an intelligent system, to supply a user of an application with personalized content so as to enhance what is refered to as "user experience" e.g. recommending a product on a merchant website or even an article on a blog. Recommendation is considered dynamic when the content to recommend or user tastes evolve rapidly e.g. news recommendation. Many applications that are of interest to us generates a tremendous amount of data through the millions of online users they have. Nevertheless, using this data to evaluate a new recommendation technique or even compare two dynamic recommendation algorithms is far from trivial. This is the problem we consider here. Some approaches have already been proposed. Nonetheless they were not studied very thoroughly both from a theoretical point of view (unquantified bias, loose convergence bounds...) and from an empirical one (experiments on private data only). In this work we start by filling many blanks within the theoretical analysis. Then we comment on the result of an experiment of unprecedented scale in this area: a public challenge we organized. This challenge along with a some complementary experiments revealed a unexpected source of a huge bias: time acceleration. The rest of this work tackles this issue. We show that a bootstrap-based approach allows to significantly reduce this…

Advisors/Committee Members: Preux, Philippe (thesis director), Mary, Jérémie (thesis director).

Subjects/Keywords: Jeux de bandits; 006.31

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Nicol, O. (2014). Data-driven evaluation of contextual bandit algorithms and applications to dynamic recommendation : Évaluation basée sur des données d'algorithmes de bandits contextuels et application à la recommandation dynamique. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2014LIL10211

Chicago Manual of Style (16^th Edition):

Nicol, Olivier. “Data-driven evaluation of contextual bandit algorithms and applications to dynamic recommendation : Évaluation basée sur des données d'algorithmes de bandits contextuels et application à la recommandation dynamique.” 2014. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2014LIL10211.

MLA Handbook (7^th Edition):

Nicol, Olivier. “Data-driven evaluation of contextual bandit algorithms and applications to dynamic recommendation : Évaluation basée sur des données d'algorithmes de bandits contextuels et application à la recommandation dynamique.” 2014. Web. 10 Apr 2021.

Vancouver:

Nicol O. Data-driven evaluation of contextual bandit algorithms and applications to dynamic recommendation : Évaluation basée sur des données d'algorithmes de bandits contextuels et application à la recommandation dynamique. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2014. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2014LIL10211.

Council of Science Editors:

Nicol O. Data-driven evaluation of contextual bandit algorithms and applications to dynamic recommendation : Évaluation basée sur des données d'algorithmes de bandits contextuels et application à la recommandation dynamique. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2014. Available from: http://www.theses.fr/2014LIL10211

8. Soare, Marta. Sequential resources allocation in linear stochastic bandits : Allocation séquentielle de ressources dans le modèle de bandit linéaire.

Degree: Docteur es, Informatique, 2015, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2015LIL10147

►

Dans cette thèse nous étudions des problèmes d'allocation de ressources dans des environnements incertains où un agent choisit ses actions séquentiellement. Après chaque pas, l'environnement… (more)

▼

Dans cette thèse nous étudions des problèmes d'allocation de ressources dans des environnements incertains où un agent choisit ses actions séquentiellement. Après chaque pas, l'environnement fournit une observation bruitée sur la valeur de l'action choisie et l'agent doit utiliser ces observations pour allouer ses ressources de façon optimale. Dans le cadre le plus classique, dit modèle du bandit à plusieurs bras (MAB), on fait l'hypothèse que chaque observation est tirée aléatoirement d'une distribution de probabilité associée à l'action choisie et ne fournit aucune information sur les valeurs espérées des autres actions disponibles dans l'environnement. Ce modèle a été largement étudié dans la littérature et plusieurs stratégies optimales ont été proposées, notamment pour le cas où le but de l'agent est de maximiser la somme des observations. Ici, nous considérons une version du MAB où les actions ne sont plus indépendantes, mais chaque observation peut être utilisée pour estimer les valeurs de l'ensemble des actions de l'environnement. Plus précisément, nous proposons des stratégies d'allocation de ressources qui sont efficaces et adaptées à un environnement caractérisé par une structure linéaire globale. Nous étudions notamment les séquences d'actions qui mènent à : (i) identifier la meilleure action avec une précision donnée et en utilisant un nombre minimum d'observations, ou (ii) maximiser la précision d'estimation des valeurs de chaque action. De plus, nous étudions les cas où les observations provenant d'un algorithme de bandit dans un environnement donné peuvent améliorer par la suite la performance de l'agent dans d'autres environnements similaires.

This thesis is dedicated to the study of resource allocation problems in uncertain environments, where an agent can sequentially select which action to take. After each step, the environment returns a noisy observation of the value of the selected action. These observations guide the agent in adapting his resource allocation strategy towards reaching a given objective. In the most typical setting of this kind, the stochastic multi-armed bandit (MAB), it is assumed that each observation is drawn from an unknown probability distribution associated with the selected action and gives no information on the expected value of the other actions. This setting has been widely studied and optimal allocation strategies were proposed to solve various objectives under the MAB assumptions. Here, we consider a variant of the MAB setting where there exists a global linear structure in the environment and by selecting an action, the agent also gathers information on the value of the other actions. Therefore, the agent needs to adapt his resource allocation strategy to exploit the structure in the environment. In particular, we study the design of sequences of actions that the agent should take to reach objectives such as: (i) identifying the best value with a fixed confidence and using a minimum number of pulls, or (ii) minimizing the prediction error on the value of each…

Advisors/Committee Members: Munos, Rémi (thesis director), Lazaric, Alessandro (thesis director).

Subjects/Keywords: Théorie de bandits; 006.31

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Soare, M. (2015). Sequential resources allocation in linear stochastic bandits : Allocation séquentielle de ressources dans le modèle de bandit linéaire. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2015LIL10147

Chicago Manual of Style (16^th Edition):

Soare, Marta. “Sequential resources allocation in linear stochastic bandits : Allocation séquentielle de ressources dans le modèle de bandit linéaire.” 2015. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2015LIL10147.

MLA Handbook (7^th Edition):

Soare, Marta. “Sequential resources allocation in linear stochastic bandits : Allocation séquentielle de ressources dans le modèle de bandit linéaire.” 2015. Web. 10 Apr 2021.

Vancouver:

Soare M. Sequential resources allocation in linear stochastic bandits : Allocation séquentielle de ressources dans le modèle de bandit linéaire. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2015. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2015LIL10147.

Council of Science Editors:

Soare M. Sequential resources allocation in linear stochastic bandits : Allocation séquentielle de ressources dans le modèle de bandit linéaire. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2015. Available from: http://www.theses.fr/2015LIL10147

9. Bubeck, Sébastien. Jeux de bandits et fondations du clustering : Bandits games and clustering foundations.

Degree: Docteur es, Mathématiques, 2010, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2010LIL10199

►

Ce travail de thèse s'inscrit dans le domaine du machine learning et concerne plus particulièrement les sous-catégories de l'optimisation stochastique, du online learning et du… (more)

Subjects/Keywords: Jeux de bandits; 519.27

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Bubeck, S. (2010). Jeux de bandits et fondations du clustering : Bandits games and clustering foundations. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2010LIL10199

Chicago Manual of Style (16^th Edition):

Bubeck, Sébastien. “Jeux de bandits et fondations du clustering : Bandits games and clustering foundations.” 2010. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2010LIL10199.

MLA Handbook (7^th Edition):

Bubeck, Sébastien. “Jeux de bandits et fondations du clustering : Bandits games and clustering foundations.” 2010. Web. 10 Apr 2021.

Vancouver:

Bubeck S. Jeux de bandits et fondations du clustering : Bandits games and clustering foundations. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2010. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2010LIL10199.

Council of Science Editors:

Bubeck S. Jeux de bandits et fondations du clustering : Bandits games and clustering foundations. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2010. Available from: http://www.theses.fr/2010LIL10199

University of Illinois – Urbana-Champaign

10. Jiang, Chong. Parametrized Stochastic Multi-armed Bandits with Binary Rewards.

Degree: MS, 1200, 2011, University of Illinois – Urbana-Champaign

URL: http://hdl.handle.net/2142/18352

► In this thesis, we consider the problem of multi-armed bandits with a large number of correlated arms. We assume that the arms have Bernoulli distributed… (more)

Subjects/Keywords: multi-armed bandits; machine learning

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Jiang, C. (2011). Parametrized Stochastic Multi-armed Bandits with Binary Rewards. (Thesis). University of Illinois – Urbana-Champaign. Retrieved from http://hdl.handle.net/2142/18352

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Jiang, Chong. “Parametrized Stochastic Multi-armed Bandits with Binary Rewards.” 2011. Thesis, University of Illinois – Urbana-Champaign. Accessed April 10, 2021. http://hdl.handle.net/2142/18352.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Jiang, Chong. “Parametrized Stochastic Multi-armed Bandits with Binary Rewards.” 2011. Web. 10 Apr 2021.

Vancouver:

Jiang C. Parametrized Stochastic Multi-armed Bandits with Binary Rewards. [Internet] [Thesis]. University of Illinois – Urbana-Champaign; 2011. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/2142/18352.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Jiang C. Parametrized Stochastic Multi-armed Bandits with Binary Rewards. [Thesis]. University of Illinois – Urbana-Champaign; 2011. Available from: http://hdl.handle.net/2142/18352

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

University of Sydney

11. Yu, Baosheng. Robust Diversity-Driven Subset Selection in Combinatorial Optimization .

Degree: 2019, University of Sydney

URL: http://hdl.handle.net/2123/19834

► Subset selection is fundamental in combinatorial optimization with applications in biology, operations research, and computer science, especially machine learning and computer vision. However, subset selection… (more)

Subjects/Keywords: subset selection; bandits; submodular optimization

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Yu, B. (2019). Robust Diversity-Driven Subset Selection in Combinatorial Optimization . (Thesis). University of Sydney. Retrieved from http://hdl.handle.net/2123/19834

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Yu, Baosheng. “Robust Diversity-Driven Subset Selection in Combinatorial Optimization .” 2019. Thesis, University of Sydney. Accessed April 10, 2021. http://hdl.handle.net/2123/19834.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Yu, Baosheng. “Robust Diversity-Driven Subset Selection in Combinatorial Optimization .” 2019. Web. 10 Apr 2021.

Vancouver:

Yu B. Robust Diversity-Driven Subset Selection in Combinatorial Optimization . [Internet] [Thesis]. University of Sydney; 2019. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/2123/19834.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Yu B. Robust Diversity-Driven Subset Selection in Combinatorial Optimization . [Thesis]. University of Sydney; 2019. Available from: http://hdl.handle.net/2123/19834

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

12. Gajane, Pratik. Multi-armed bandits with unconventional feedback : Bandits multi-armés avec rétroaction partielle.

Degree: Docteur es, Informatique, 2017, Lille 3

URL: http://www.theses.fr/2017LIL30045

►

Dans cette thèse, nous étudions des problèmes de prise de décisions séquentielles dans lesquels, pour chacune de ses décisions, l'apprenant reçoit une information qu'il utilise… (more)

▼

Dans cette thèse, nous étudions des problèmes de prise de décisions séquentielles dans lesquels, pour chacune de ses décisions, l'apprenant reçoit une information qu'il utilise pour guider ses décisions futures. Pour aller au-delà du retour d’information conventionnel tel qu'il a été bien étudié pour des problèmes de prise de décision séquentielle tels que les bandits multi-bras, nous considérons des formes de retour d’information partielle motivées par des applications pratiques.En premier, nous considérons le problème des bandits duellistes, dans lequel l'apprenant sélectionne deux actions à chaque pas de temps et reçoit en retour une information relative (i.e. de préférence) entre les valeurs instantanées de ces deux actions.En particulier, nous proposons un algorithme optimal qui permet à l'apprenant d'obtenir un regret cumulatif quasi-optimal (le regret est la différence entre la récompense cumulative optimale et la récompense cumulative constatée de l’apprenant). Dans un second temps, nous considérons le problème des bandits corrompus, dans lequel un processus de corruption stochastique perturbe le retour d’information. Pour ce problème aussi, nous concevons des algorithmes pour obtenir un regret cumulatif asymptotiquement optimal. En outre, nous examinons la relation entre ces deux problèmes dans le cadre du monitoring partiel qui est un paradigme générique pour la prise de décision séquentielle avec retour d'information partielle.

The multi-armed bandit (MAB) problem is a mathematical formulation of the exploration-exploitation trade-off inherent to reinforcement learning, in which the learner chooses an action (symbolized by an arm) from a set of available actions in a sequence of trials in order to maximize their reward. In the classical MAB problem, the learner receives absolute bandit feedback i.e. it receives as feedback the reward of the arm it selects. In many practical situations however, different kind of feedback is more readily available. In this thesis, we study two of such kinds of feedbacks, namely, relative feedback and corrupt feedback.The main practical motivation behind relative feedback arises from the task of online ranker evaluation. This task involves choosing the optimal ranker from a finite set of rankers using only pairwise comparisons, while minimizing the comparisons between sub-optimal rankers. This is formalized by the MAB problem with relative feedback, in which the learner selects two arms instead of one and receives the preference feedback. We consider the adversarial formulation of this problem which circumvents the stationarity assumption over the mean rewards for the arms. We provide a lower bound on the performance measure for any algorithm for this problem. We also provide an algorithm called "Relative Exponential-weight algorithm for Exploration and Exploitation" with performance guarantees. We present a thorough empirical study on several information retrieval datasets that confirm the validity of these theoretical results.The motivating theme behind corrupt feedback…

Advisors/Committee Members: Preux, Philippe (thesis director).

Subjects/Keywords: Bandits Multi-Bras; Retour D’information Partielle; Dueling Bandits; Corrupt Bandits; Évaluation du Ranker; Vie Privée Différentielle; Multi-Armed Bandit; Partial Feedback; Dueling Bandits; Corrupt Bandits; Ranker Evaluation; Differential Privacy

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Gajane, P. (2017). Multi-armed bandits with unconventional feedback : Bandits multi-armés avec rétroaction partielle. (Doctoral Dissertation). Lille 3. Retrieved from http://www.theses.fr/2017LIL30045

Chicago Manual of Style (16^th Edition):

Gajane, Pratik. “Multi-armed bandits with unconventional feedback : Bandits multi-armés avec rétroaction partielle.” 2017. Doctoral Dissertation, Lille 3. Accessed April 10, 2021. http://www.theses.fr/2017LIL30045.

MLA Handbook (7^th Edition):

Gajane, Pratik. “Multi-armed bandits with unconventional feedback : Bandits multi-armés avec rétroaction partielle.” 2017. Web. 10 Apr 2021.

Vancouver:

Gajane P. Multi-armed bandits with unconventional feedback : Bandits multi-armés avec rétroaction partielle. [Internet] [Doctoral dissertation]. Lille 3; 2017. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2017LIL30045.

Council of Science Editors:

Gajane P. Multi-armed bandits with unconventional feedback : Bandits multi-armés avec rétroaction partielle. [Doctoral Dissertation]. Lille 3; 2017. Available from: http://www.theses.fr/2017LIL30045

University of South Carolina

13. Dawson, Tracie Renée. Elsewhere, Then.

Degree: MFA, Art, 2017, University of South Carolina

URL: https://scholarcommons.sc.edu/etd/4193

► Elsewhere, Then is a cross-generational road novel that traces the sociopolitical and personal narrative of a family through a musician making her way from… (more)

Subjects/Keywords: Arts and Humanities; Creative Writing; Elsewhere; Then

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Dawson, T. R. (2017). Elsewhere, Then. (Thesis). University of South Carolina. Retrieved from https://scholarcommons.sc.edu/etd/4193

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Dawson, Tracie Renée. “Elsewhere, Then.” 2017. Thesis, University of South Carolina. Accessed April 10, 2021. https://scholarcommons.sc.edu/etd/4193.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Dawson, Tracie Renée. “Elsewhere, Then.” 2017. Web. 10 Apr 2021.

Vancouver:

Dawson TR. Elsewhere, Then. [Internet] [Thesis]. University of South Carolina; 2017. [cited 2021 Apr 10]. Available from: https://scholarcommons.sc.edu/etd/4193.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Dawson TR. Elsewhere, Then. [Thesis]. University of South Carolina; 2017. Available from: https://scholarcommons.sc.edu/etd/4193

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

University of Georgia

14. Sebastian, Bram. Computational prediction of chicken pre-microRNA.

Degree: 2014, University of Georgia

URL: http://hdl.handle.net/10724/24488

► MicroRNAs (miRNAs) are a small single strand non-coding RNA with ~22 nucleotides (nts) which can regulate gene expression. MiRNAs are generated from a ~60-70 nts… (more)

Subjects/Keywords: computational prediction; microRNA; chicken; miR-Explore

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Sebastian, B. (2014). Computational prediction of chicken pre-microRNA. (Thesis). University of Georgia. Retrieved from http://hdl.handle.net/10724/24488

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Sebastian, Bram. “Computational prediction of chicken pre-microRNA.” 2014. Thesis, University of Georgia. Accessed April 10, 2021. http://hdl.handle.net/10724/24488.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Sebastian, Bram. “Computational prediction of chicken pre-microRNA.” 2014. Web. 10 Apr 2021.

Vancouver:

Sebastian B. Computational prediction of chicken pre-microRNA. [Internet] [Thesis]. University of Georgia; 2014. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/10724/24488.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Sebastian B. Computational prediction of chicken pre-microRNA. [Thesis]. University of Georgia; 2014. Available from: http://hdl.handle.net/10724/24488

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

15. Gabillon, Victor. Algorithmes budgétisés d'itérations sur les politiques obtenues par classification : Budgeted classification-based policy iteration.

Degree: Docteur es, Informatique, 2014, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2014LIL10032

►

Cette thèse étudie une classe d'algorithmes d'apprentissage par renforcement (RL), appelée « itération sur les politiques obtenues par classification » (CBPI). Contrairement aux méthodes standards… (more)

▼

Cette thèse étudie une classe d'algorithmes d'apprentissage par renforcement (RL), appelée « itération sur les politiques obtenues par classification » (CBPI). Contrairement aux méthodes standards de RL, CBPI n'utilise pas de représentation explicite de la fonction valeur. CBPI réalise des déroulés (des trajectoires) et estime la fonction action-valeur de la politique courante pour un nombre limité d'états et d'actions. En utilisant un ensemble d'apprentissage construit à partir de ces estimations, la politique gloutonne est apprise comme le produit d'un classificateur. La politique ainsi produite à chaque itération de l'algorithme, n'est plus définie par une fonction valeur (approximée), mais par un classificateur. Dans cette thèse, nous proposons de nouveaux algorithmes qui améliorent les performances des méthodes CBPI existantes, spécialement lorsque le nombre d’interactions avec l’environnement est limité. Nos améliorations se portent sur les deux limitations de CBPI suivantes : 1) les déroulés utilisés pour estimer les fonctions action-valeur doivent être tronqués et leur nombre est limité, créant un compromis entre le biais et la variance dans ces estimations, et 2) les déroulés sont répartis de manière uniforme entre les états déroulés et les actions disponibles, alors qu'une stratégie plus évoluée pourrait garantir un ensemble d'apprentissage plus précis. Nous proposons des algorithmes CBPI qui répondent à ces limitations, respectivement : 1) en utilisant une approximation de la fonction valeur pour améliorer la précision (en équilibrant biais et variance) des estimations, et 2) en échantillonnant de manière adaptative les déroulés parmi les paires d'état-action.

This dissertation is motivated by the study of a class of reinforcement learning (RL) algorithms, called classification-based policy iteration (CBPI). Contrary to the standard RL methods, CBPI do not use an explicit representation for value function. Instead, they use rollouts and estimate the action-value function of the current policy at a collection of states. Using a training set built from these rollout estimates, the greedy policy is learned as the output of a classifier. Thus, the policy generated at each iteration of the algorithm, is no longer defined by a (approximated) value function, but instead by a classifier. In this thesis, we propose new algorithms that improve the performance of the existing CBPI methods, especially when they have a fixed budget of interaction with the environment. Our improvements are based on the following two shortcomings of the existing CBPI algorithms: 1) The rollouts that are used to estimate the action-value functions should be truncated and their number is limited, and thus, we have to deal with bias-variance tradeoff in estimating the rollouts, and 2) The rollouts are allocated uniformly over the states in the rollout set and the available actions, while a smarter allocation strategy could guarantee a more accurate training set for the classifier. We propose CBPI algorithms that address these issues,…

Advisors/Committee Members: Preux, Philippe (thesis director), Ghavamzadeh, Mohammad (thesis director).

Subjects/Keywords: Jeux de bandits; Apprentissage par renforcement; 006.31

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Gabillon, V. (2014). Algorithmes budgétisés d'itérations sur les politiques obtenues par classification : Budgeted classification-based policy iteration. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2014LIL10032

Chicago Manual of Style (16^th Edition):

Gabillon, Victor. “Algorithmes budgétisés d'itérations sur les politiques obtenues par classification : Budgeted classification-based policy iteration.” 2014. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2014LIL10032.

MLA Handbook (7^th Edition):

Gabillon, Victor. “Algorithmes budgétisés d'itérations sur les politiques obtenues par classification : Budgeted classification-based policy iteration.” 2014. Web. 10 Apr 2021.

Vancouver:

Gabillon V. Algorithmes budgétisés d'itérations sur les politiques obtenues par classification : Budgeted classification-based policy iteration. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2014. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2014LIL10032.

Council of Science Editors:

Gabillon V. Algorithmes budgétisés d'itérations sur les politiques obtenues par classification : Budgeted classification-based policy iteration. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2014. Available from: http://www.theses.fr/2014LIL10032

16. Kocák, Tomáš. Apprentissage séquentiel avec similitudes : Sequential learning with similarities.

Degree: Docteur es, Informatique, 2016, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2016LIL10230

►

Dans cette thèse nous étudions différentes généralisations du problème dit « du bandit manchot ». Le problème du bandit manchot est un problème de décision… (more)

▼

Dans cette thèse nous étudions différentes généralisations du problème dit « du bandit manchot ». Le problème du bandit manchot est un problème de décision séquentiel au cours duquel un agent sélectionne successivement des actions et obtient une récompense pour chacune d'elles. On fait généralement l'hypothèse que seule la récompense associée à l'action choisie est observée par l'agent, ce dernier ne reçoit aucune information sur les actions non choisies. Cette hypothèse s'avère parfois très restrictive pour certains problèmes très structurés tels que les systèmes de recommandations, la publicité en ligne, le routage de paquets, etc. Il paraît assez naturel de tenir compte de la connaissance de la structure du problème pour améliorer les performances des algorithmes d'apprentissage usuels. Dans cette thèse, nous nous focalisons sur les problèmes de bandits présentant une structure pouvant être modélisée par un graphe dont les nœuds représentent les actions. Dans un premier temps, nous étudierons le cas où les arêtes du graphe modélisent les similitudes entre actions. Dans un second temps, nous analyserons le cas où l'agent observe les récompenses de toutes les actions adjacentes à l'action choisie dans le graphe. Notre contribution principale a été d'élaborer de nouveaux algorithmes permettant de traiter efficacement les problèmes évoqués précédemment, et de démontrer théoriquement et empiriquement le bon fonctionnement de ces algorithmes. Nos travaux nous ont également amenés à introduire de nouvelles grandeurs, telles que la dimension effective et le nombre d'indépendance effectif, afin de caractériser la difficulté des différents problèmes.

This thesis studies several extensions of multi-armed bandit problem, where a learner sequentially selects an action and obtain the reward of the action. Traditionally, the only information the learner acquire is about the obtained reward while information about other actions is hidden from the learner. This limited feedback can be restrictive in some applications like recommender systems, internet advertising, packet routing, etc. Usually, these problems come with structure, similarities between users or actions, additional observations, or any additional assumptions. Therefore, it is natural to incorporate these assumptions to the algorithms to improve their performance. This thesis focuses on multi-armed bandit problem with some underlying structure usually represented by a graph with actions as vertices. First, we study a problem where the graph captures similarities between actions; connected actions tend to grand similar rewards. Second, we study a problem where the learner observes rewards of all the neighbors of the selected action. We study these problems under several additional assumptions on rewards (stochastic, adversarial), side observations (adversarial, stochastic, noisy), actions (one node at the time, several nodes forming a combinatorial structure in the graph). The main contribution of this thesis is to design algorithms for previously mentioned problems…

Advisors/Committee Members: Valko, Michal (thesis director), Munos, Rémi (thesis director).

Subjects/Keywords: Jeux de bandits; Apprentissage séquentiel; 006.31

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Kocák, T. (2016). Apprentissage séquentiel avec similitudes : Sequential learning with similarities. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2016LIL10230

Chicago Manual of Style (16^th Edition):

Kocák, Tomáš. “Apprentissage séquentiel avec similitudes : Sequential learning with similarities.” 2016. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2016LIL10230.

MLA Handbook (7^th Edition):

Kocák, Tomáš. “Apprentissage séquentiel avec similitudes : Sequential learning with similarities.” 2016. Web. 10 Apr 2021.

Vancouver:

Kocák T. Apprentissage séquentiel avec similitudes : Sequential learning with similarities. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2016. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2016LIL10230.

Council of Science Editors:

Kocák T. Apprentissage séquentiel avec similitudes : Sequential learning with similarities. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2016. Available from: http://www.theses.fr/2016LIL10230

17. Maillard, Odalric-Ambrym. Apprentissage séquentiel : bandits, statistique et renforcement : Sequential Learning : Bandits, Statistics and Reinforcement.

Degree: Docteur es, Informatique, 2011, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2011LIL10041

►

Cette thèse traite des domaines suivant en Apprentissage Automatique: la théorie des Bandits, l'Apprentissage statistique et l'Apprentissage par renforcement. Son fil rouge est l'étude de… (more)

▼

Cette thèse traite des domaines suivant en Apprentissage Automatique: la théorie des Bandits, l'Apprentissage statistique et l'Apprentissage par renforcement. Son fil rouge est l'étude de plusieurs notions d'adaptation, d'un point de vue non asymptotique : à un environnement ou à un adversaire dans la partie I, à la structure d'un signal dans la partie II, à la structure de récompenses ou à un modèle des états du monde dans la partie III. Tout d'abord nous dérivons une analyse non asymptotique d'un algorithme de bandit à plusieurs bras utilisant la divergence de Kullback-Leibler. Celle-ci permet d'atteindre, dans le cas de distributions à support fini, la borne inférieure de performance asymptotique dépendante des distributions de probabilité connue pour ce problème. Puis, pour un bandit avec un adversaire possiblement adaptatif, nous introduisons des modèles dépendants de l'histoire et traduisant une possible faiblesse de l'adversaire et montrons comment en tirer parti pour concevoir des algorithmes adaptatifs à cette faiblesse. Nous contribuons au problème de la régression en montrant l'utilité des projections aléatoires, à la fois sur le plan théorique et pratique, lorsque l'espace d'hypothèses considéré est de dimension grande, voire infinie. Nous utilisons également des opérateurs d'échantillonnage aléatoires dans le cadre de la reconstruction parcimonieuse lorsque la base est loin d'être orthogonale. Enfin, nous combinons la partie I et II : pour fournir une analyse non-asymptotique d'algorithmes d'apprentissage par renforcement; puis, en amont du cadre des Processus Décisionnel de Markov, pour discuter du problème pratique du choix d'un bon modèle d'états.

This thesis studies the following topics in Machine Learning: Bandit theory, Statistical learning and Reinforcement learning. The common underlying thread is the non-asymptotic study of various notions of adaptation : to an environment or an opponent in part I about bandit theory, to the structure of a signal in part II about statistical theory, to the structure of states and rewards or to some state-model of the world in part III about reinforcement learning. First we derive a non-asymptotic analysis of a Kullback-Leibler-based algorithm for the stochastic multi-armed bandit that enables to match, in the case of distributions with finite support, the asymptotic distribution-dependent lower bound known for this problem. Now for a multi-armed bandit with a possibly adaptive opponent, we introduce history-based models to catch some weakness of the opponent, and show how one can benefit from such models to design algorithms adaptive to this weakness. Then we contribute to the regression setting and show how the use of random matrices can be beneficial both theoretically and numerically when the considered hypothesis space has a large, possibly infinite, dimension. We also use random matrices in the sparse recovery setting to build sensing operators that allow for recovery when the basis is far from being orthogonal. Finally we combine part I and II to first…

Advisors/Committee Members: Munos, Rémi (thesis director), Berthet, Philippe (thesis director).

Subjects/Keywords: Jeux de bandits; Apprentissage par renforcement

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Maillard, O. (2011). Apprentissage séquentiel : bandits, statistique et renforcement : Sequential Learning : Bandits, Statistics and Reinforcement. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2011LIL10041

Chicago Manual of Style (16^th Edition):

Maillard, Odalric-Ambrym. “Apprentissage séquentiel : bandits, statistique et renforcement : Sequential Learning : Bandits, Statistics and Reinforcement.” 2011. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2011LIL10041.

MLA Handbook (7^th Edition):

Maillard, Odalric-Ambrym. “Apprentissage séquentiel : bandits, statistique et renforcement : Sequential Learning : Bandits, Statistics and Reinforcement.” 2011. Web. 10 Apr 2021.

Vancouver:

Maillard O. Apprentissage séquentiel : bandits, statistique et renforcement : Sequential Learning : Bandits, Statistics and Reinforcement. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2011. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2011LIL10041.

Council of Science Editors:

Maillard O. Apprentissage séquentiel : bandits, statistique et renforcement : Sequential Learning : Bandits, Statistics and Reinforcement. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2011. Available from: http://www.theses.fr/2011LIL10041

18. Carpentier, Alexandra. De l'échantillonage optimal en grande et petite dimension : On optimal sampling in high and low dimension.

Degree: Docteur es, Informatique, 2012, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2012LIL10041

►

Pendant ma thèse, j’ai eu la chance d’apprendre et de travailler sous la supervision de mon directeur de thèse Rémi, et ce dans deux domaines… (more)

Subjects/Keywords: Théorie des bandits stochastiques; Compressed sensing; 006.31

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Carpentier, A. (2012). De l'échantillonage optimal en grande et petite dimension : On optimal sampling in high and low dimension. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2012LIL10041

Chicago Manual of Style (16^th Edition):

Carpentier, Alexandra. “De l'échantillonage optimal en grande et petite dimension : On optimal sampling in high and low dimension.” 2012. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2012LIL10041.

MLA Handbook (7^th Edition):

Carpentier, Alexandra. “De l'échantillonage optimal en grande et petite dimension : On optimal sampling in high and low dimension.” 2012. Web. 10 Apr 2021.

Vancouver:

Carpentier A. De l'échantillonage optimal en grande et petite dimension : On optimal sampling in high and low dimension. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2012. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2012LIL10041.

Council of Science Editors:

Carpentier A. De l'échantillonage optimal en grande et petite dimension : On optimal sampling in high and low dimension. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2012. Available from: http://www.theses.fr/2012LIL10041

University of Illinois – Urbana-Champaign

19. Magesh, Akshayaa. Decentralized multi-user multi-armed bandits with user dependent reward distributions.

Degree: MS, Electrical & Computer Engr, 2020, University of Illinois – Urbana-Champaign

URL: http://hdl.handle.net/2142/107917

► The uncoordinated spectrum access problem is studied using a multi-player multi-armed bandits framework. We consider a decentralized multi-player stochastic multi-armed bandit model where the players… (more)

Subjects/Keywords: multiarmed bandits; multi-player; spectrum access; decentralized

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Magesh, A. (2020). Decentralized multi-user multi-armed bandits with user dependent reward distributions. (Thesis). University of Illinois – Urbana-Champaign. Retrieved from http://hdl.handle.net/2142/107917

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Magesh, Akshayaa. “Decentralized multi-user multi-armed bandits with user dependent reward distributions.” 2020. Thesis, University of Illinois – Urbana-Champaign. Accessed April 10, 2021. http://hdl.handle.net/2142/107917.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Magesh, Akshayaa. “Decentralized multi-user multi-armed bandits with user dependent reward distributions.” 2020. Web. 10 Apr 2021.

Vancouver:

Magesh A. Decentralized multi-user multi-armed bandits with user dependent reward distributions. [Internet] [Thesis]. University of Illinois – Urbana-Champaign; 2020. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/2142/107917.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Magesh A. Decentralized multi-user multi-armed bandits with user dependent reward distributions. [Thesis]. University of Illinois – Urbana-Champaign; 2020. Available from: http://hdl.handle.net/2142/107917

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

University of Illinois – Urbana-Champaign

20. Jiang, Chong. Online advertisements and multi-armed bandits.

Degree: PhD, Electrical & Computer Engr, 2015, University of Illinois – Urbana-Champaign

URL: http://hdl.handle.net/2142/78369

► We investigate a number of multi-armed bandit problems that model different aspects of online advertising, beginning with a survey of the key techniques that are… (more)

▼ We investigate a number of multi-armed bandit problems that model different aspects of online advertising, beginning with a survey of the key techniques that are commonly used to demonstrate the theoretical limitations and achievable results for the performance of multi-armed bandit algorithms. We then formulate variations of the basic stochastic multi-armed bandit problem, aimed at modeling how budget-limited advertisers should bid and how ad exchanges should choose whose ad to display, and study them using these techniques. We first consider online ad auctions from the point of view of a single advertiser who has an average budget constraint. By modeling the rest of the bidders through a probability distribution (often referred to as the mean-field approximation), we develop a simple bidding strategy which can be implemented without any statistical knowledge of bids, valuations, and query arrival processes. The key idea is to use stochastic approximation techniques to automatically track long-term averages. Next, we consider multi-armed bandits with budgets, modeling how ad exchanges select which ad to display. We provide asymptotic regret lower bounds satisfied by any algorithm, and propose algorithms which match those lower bounds. We consider different types of budgets: scenarios where the advertiser has a fixed budget over a time horizon, and scenarios where the amount of money that is available to spend is incremented in each time slot. Further, we consider two different pricing models, one in which an advertiser is charged each time their ad is shown, and one in which the advertiser is charged only if a user clicks on the ad. For all of these cases, we show that it is possible to achieve O(log(T)) regret. For both the cost-per-impression and cost-per-click models, with a fixed budget, we provide regret lower bounds that apply to any uniformly good algorithm. Further, we show that B-KL-UCB, a natural variant of KL-UCB, is asymptotically optimal for these cases. Numerical experiments (based on a real-world data set) further suggest that B-KL-UCB also has the same or better finite-time performance when compared to various previously proposed (UCB-like) algorithms. Finally, we consider the problem of multi-armed bandits with a large, possibly infinite number of correlated arms, modeling a retailer advertising a large number of related items. We assume that the arms have Bernoulli distributed rewards, where the probabilities of success are parametrized by known attribute vectors for each arm and an unknown vector which describes the preferences of the target audience. For this model, we seek an algorithm with a total regret that is sub-linear in time and independent of the number of arms. We present such an algorithm and analyze its performance, showing upper bounds on the total regret which apply uniformly in time, for both the finite and infinite arm cases. Advisors/Committee Members: Srikant, R. (advisor), Srikant, R. (Committee Chair), Beck, Carolyn (committee member), Nedich, Angelia (committee member), Veeravalli, Venugopal V. (committee member).

Subjects/Keywords: Multi-armed bandits; online advertisements; reinforcement learning

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Jiang, C. (2015). Online advertisements and multi-armed bandits. (Doctoral Dissertation). University of Illinois – Urbana-Champaign. Retrieved from http://hdl.handle.net/2142/78369

Chicago Manual of Style (16^th Edition):

Jiang, Chong. “Online advertisements and multi-armed bandits.” 2015. Doctoral Dissertation, University of Illinois – Urbana-Champaign. Accessed April 10, 2021. http://hdl.handle.net/2142/78369.

MLA Handbook (7^th Edition):

Jiang, Chong. “Online advertisements and multi-armed bandits.” 2015. Web. 10 Apr 2021.

Vancouver:

Jiang C. Online advertisements and multi-armed bandits. [Internet] [Doctoral dissertation]. University of Illinois – Urbana-Champaign; 2015. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/2142/78369.

Council of Science Editors:

Jiang C. Online advertisements and multi-armed bandits. [Doctoral Dissertation]. University of Illinois – Urbana-Champaign; 2015. Available from: http://hdl.handle.net/2142/78369

University of Minnesota

21. Beek, Aaron. Freelance Warfare and Illegitimacy: the Historians’ Portrayal of Bandits, Pirates, Mercenaries and Politicians.

Degree: PhD, Classical and Near Eastern Studies, 2015, University of Minnesota

URL: http://hdl.handle.net/11299/174839

► This dissertation examines freelance warfare in the ancient world. The "freelancer" needs to be understood as a unified category, not compartmentalized as three (or more)… (more)

▼ This dissertation examines freelance warfare in the ancient world. The "freelancer" needs to be understood as a unified category, not compartmentalized as three (or more) groups: pirates, bandits, and mercenaries. Throughout, I contend that ancient authors' perception and portrayal of the actions of freelancers dramatically affected the perceived legitimacy of those actions. Most other studies (e.g. Shaw 1984, de Souza 1999, Grünewald 1999, Pohl 1993, Trundle 2004, Knapp 2011) focus on 'real' bandits and on a single one of these groups. I examine these three groups together, but also ask what semantic baggage words like latro or leistes had to carry that they were commonly used in invectives. Thus rhetorical piracy is also important for my study. The work unfolds in three parts. The first is a brief chronological survey of "freelance men of violence" of all stripes down to the second century BC. Freelancers engage in, at best, semi-legitimate acts of force. Excluded are standing paid forces and theft by means other than force, vis. In a form of ancient realpolitik, the freelancer was generally more acceptable to states than our aristocratic historians would prefer that we believe. Moreover, states were more concerned with control of these "freelancers" than in their elimination. The second section explains events of the second and first century in greater detail. The observations made in the first section hold true in the second, despite being depicted differently by ancient historians. The third section focuses on the historians, historical accounts and rhetoric employed. The historians make motivations less pragmatic and more idealistic. Additionally, the perception of piracy was affected by triumphal politics, consular authority, and employment of mercenaries Overall, the chief semantic burden of pirate-terms is to convey legitimacy: individuals that possess power that they should not. Condemnation of these figures is not rooted in their actions of plundering (rarely dissimilar from official acts of war) but instead their holding any such power in the first place. In short, this study reveals that the "at-large" soldier was far more complex and far more influential than is normally shown by either ancient or modern historians.

Subjects/Keywords: Bandits; Freelancers; Legitimacy; Mercenaries; Pirates; Portrayal

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Beek, A. (2015). Freelance Warfare and Illegitimacy: the Historians’ Portrayal of Bandits, Pirates, Mercenaries and Politicians. (Doctoral Dissertation). University of Minnesota. Retrieved from http://hdl.handle.net/11299/174839

Chicago Manual of Style (16^th Edition):

Beek, Aaron. “Freelance Warfare and Illegitimacy: the Historians’ Portrayal of Bandits, Pirates, Mercenaries and Politicians.” 2015. Doctoral Dissertation, University of Minnesota. Accessed April 10, 2021. http://hdl.handle.net/11299/174839.

MLA Handbook (7^th Edition):

Beek, Aaron. “Freelance Warfare and Illegitimacy: the Historians’ Portrayal of Bandits, Pirates, Mercenaries and Politicians.” 2015. Web. 10 Apr 2021.

Vancouver:

Beek A. Freelance Warfare and Illegitimacy: the Historians’ Portrayal of Bandits, Pirates, Mercenaries and Politicians. [Internet] [Doctoral dissertation]. University of Minnesota; 2015. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/11299/174839.

Council of Science Editors:

Beek A. Freelance Warfare and Illegitimacy: the Historians’ Portrayal of Bandits, Pirates, Mercenaries and Politicians. [Doctoral Dissertation]. University of Minnesota; 2015. Available from: http://hdl.handle.net/11299/174839

University of Minnesota

22. Sivakumar, Vidyashankar. Beyond Sub-Gaussian and Independent Data in High Dimensional Regression.

Degree: PhD, Computer Science, 2020, University of Minnesota

URL: http://hdl.handle.net/11299/217800

► The past three decades has seen major developments in high-dimensional regression models leading to their successful use in applications from multiple domains including climate science,… (more)

▼ The past three decades has seen major developments in high-dimensional regression models leading to their successful use in applications from multiple domains including climate science, finance, recommendation systems, computational biology, signal processing to name a few. The underlying assumption in high-dimensional regression models is that the phenomenon under study can be explained by a simple model with few variables. In high-dimensional parametric regression models with parameters existing in high-dimensional space, the simplicity assumption is encoded by a sparsity constraint to be satisfied by the parameter vector. Statistical analysis of high-dimensional regression models delves into the study of the properties of the models, including how faithfully the models recover the assumed true sparse parameter and model sensitivity to different data assumptions. While major progress has been made over the past several years, non-asymptotic statistical analysis of high-dimensional regression models still makes standard data assumptions of (sub)-Gaussianity and independence which do not hold in some practical applications. For example, datasets in climate and finance are known to have variables with heavier tails than Gaussian or bandit algorithms have data that is sequentially chosen thus violating the independence assumption. The topic of this thesis is the non-asymptotic statistical analysis and study of high-dimensional regression estimators under non-standard data assumptions, including analysis of traditional estimators like regularized least squares as also design of new algorithms to improve estimation performance. Our technical results highlight geometric properties of high-dimensional models and hence all results are expressed in terms of geometric quantities associated with the sparsity structure assumed for the parameter. Much of the analysis borrows tools and techniques from random matrix analysis, probability tools like generic chaining and, in general, probability results for behavior of random variables, vectors in high-dimensional space. We analyze four problems: 1. Regularized least squares with sub-exponential data: Data in multiple domains like finance, climate science are known to be sub-exponential, which have probability distributions with tails heavier than Gaussians but dominated by a suitably scaled centered exponential distribution. We study non-asymptotic estimation performance of the regularized least squares estimator with i.i.d. sub-exponential data showing that the estimation performance is slightly worse compared to the i.i.d. sub-Gaussian setting. 2. High-dimensional quantile regression: We study the quantile regression problem in high dimensions which models the conditional quantile of a response given covariates. While least squares regression is ideal to model the conditional mean of a response variable which is symmetric (sub)-Gaussian, there are multiple applications where it is imperative/of interest to model conditional quantiles of the response given covariates to…

Subjects/Keywords: Bandits and online learning; High-dimensional regression

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Sivakumar, V. (2020). Beyond Sub-Gaussian and Independent Data in High Dimensional Regression. (Doctoral Dissertation). University of Minnesota. Retrieved from http://hdl.handle.net/11299/217800

Chicago Manual of Style (16^th Edition):

Sivakumar, Vidyashankar. “Beyond Sub-Gaussian and Independent Data in High Dimensional Regression.” 2020. Doctoral Dissertation, University of Minnesota. Accessed April 10, 2021. http://hdl.handle.net/11299/217800.

MLA Handbook (7^th Edition):

Sivakumar, Vidyashankar. “Beyond Sub-Gaussian and Independent Data in High Dimensional Regression.” 2020. Web. 10 Apr 2021.

Vancouver:

Sivakumar V. Beyond Sub-Gaussian and Independent Data in High Dimensional Regression. [Internet] [Doctoral dissertation]. University of Minnesota; 2020. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/11299/217800.

Council of Science Editors:

Sivakumar V. Beyond Sub-Gaussian and Independent Data in High Dimensional Regression. [Doctoral Dissertation]. University of Minnesota; 2020. Available from: http://hdl.handle.net/11299/217800

23. Abeille, Marc. Exploration-exploitation with Thompson sampling in linear systems : Algorithmes de Thompson sampling pour l’exploration-exploitation dans les systèmes linéaires.

Degree: Docteur es, Mathématiques appliquées, 2017, Université Lille I – Sciences et Technologies

URL: http://www.theses.fr/2017LIL10182

►

Cette thèse est dédiée à l'étude du Thompson Sampling (TS), une heuristique qui vise à surmonter le dilemme entre exploration et exploitation qui est inhérent… (more)

▼

Cette thèse est dédiée à l'étude du Thompson Sampling (TS), une heuristique qui vise à surmonter le dilemme entre exploration et exploitation qui est inhérent à tout processus décisionnel face à l'incertain. Contrairement aux algorithmes issus de l'heuristique optimiste face à l'incertain (OFU), où l'exploration provient du choix du modèle le plus favorable possible au vu de la connaissance accumulée, les algorithmes TS introduisent de l'aléa dans le processus décisionnel en sélectionnant aléatoirement un modèle plausible, ce qui les rend bien moins coûteux numériquement. Cette étude se concentre sur les problèmes paramétriques linéaires, qui autorisent les espaces état-action continus (infinis), en particulier les problèmes de Bandits Linéaires (LB) et les problèmes de contrôle Linéaire et Quadratique (LQ). Nous proposons dans cette thèse de nouvelles analyses du regret des algorithmes TS pour chacun de ces deux problèmes. Bien que notre démonstration pour les LB garantisse une borne supérieure identique aux résultats préexistants, la structure de la preuve offre une nouvelle vision du fonctionnement de l'algorithme TS, et nous permet d'étendre cette analyse aux problèmes LQ. Nous démontrons la première borne supérieure pour le regret de l'algorithme TS dans les problèmes LQ, qui garantie dans le cadre fréquentiste un regret au plus d'ordre O(√{T}). Enfin, nous proposons une application des méthodes d'exploration-exploitation pour les problèmes d'optimisation de portefeuille, et discutons dans ce cadre le besoin ou non d'explorer activement.

This dissertation is dedicated to the study of the Thompson Sampling (TS) algorithms designed to address the exploration-exploitation dilemma that is inherent in sequential decision-making under uncertainty. As opposed to algorithms derived from the optimism-in-the-face-of-uncertainty (OFU) principle, where the exploration is performed by selecting the most favorable model within the set of plausible one, TS algorithms rely on randomization to enhance the exploration, and thus are much more computationally efficient. We focus on linearly parametrized problems that allow for continuous state-action spaces, namely the Linear Bandit (LB) problems and the Linear Quadratic (LQ) control problems. We derive two novel analyses for the regret of TS algorithms in those settings. While the obtained regret bound for LB is similar to previous results, the proof sheds new light on the functioning of TS, and allows us to extend the analysis to LQ problems. As a result, we prove the first regret bound for TS in LQ, and show that the frequentist regret is of order O(sqrt{T}) which matches the existing guarantee for the regret of OFU algorithms in LQ. Finally, we propose an application of exploration-exploitation techniques to the practical problem of portfolio construction, and discuss the need for active exploration in this setting.

Advisors/Committee Members: Munos, Rémi (thesis director), Lazaric, Alessandro (thesis director), Sérié, Emmanuel (thesis director).

Subjects/Keywords: Algorithme Thompson sampling; Bandits multi-Bras; Bandits linéaires; Contrôle des systèmes linéaires quadratiques; 519.62

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Abeille, M. (2017). Exploration-exploitation with Thompson sampling in linear systems : Algorithmes de Thompson sampling pour l’exploration-exploitation dans les systèmes linéaires. (Doctoral Dissertation). Université Lille I – Sciences et Technologies. Retrieved from http://www.theses.fr/2017LIL10182

Chicago Manual of Style (16^th Edition):

Abeille, Marc. “Exploration-exploitation with Thompson sampling in linear systems : Algorithmes de Thompson sampling pour l’exploration-exploitation dans les systèmes linéaires.” 2017. Doctoral Dissertation, Université Lille I – Sciences et Technologies. Accessed April 10, 2021. http://www.theses.fr/2017LIL10182.

MLA Handbook (7^th Edition):

Abeille, Marc. “Exploration-exploitation with Thompson sampling in linear systems : Algorithmes de Thompson sampling pour l’exploration-exploitation dans les systèmes linéaires.” 2017. Web. 10 Apr 2021.

Vancouver:

Abeille M. Exploration-exploitation with Thompson sampling in linear systems : Algorithmes de Thompson sampling pour l’exploration-exploitation dans les systèmes linéaires. [Internet] [Doctoral dissertation]. Université Lille I – Sciences et Technologies; 2017. [cited 2021 Apr 10]. Available from: http://www.theses.fr/2017LIL10182.

Council of Science Editors:

Abeille M. Exploration-exploitation with Thompson sampling in linear systems : Algorithmes de Thompson sampling pour l’exploration-exploitation dans les systèmes linéaires. [Doctoral Dissertation]. Université Lille I – Sciences et Technologies; 2017. Available from: http://www.theses.fr/2017LIL10182

24. -4677-643X. Online experiment design with causal structures.

Degree: PhD, Electrical and Computer Engineering, 2019, University of Texas – Austin

URL: http://dx.doi.org/10.26153/tsw/2950

► Modern learning systems like recommendation engines, computational advertising systems, online parameter tuning services are inherently online; i.e. these systems need to continually collect data, take… (more)

▼ Modern learning systems like recommendation engines, computational advertising systems, online parameter tuning services are inherently online; i.e. these systems need to continually collect data, take decisions to optimize a certain objective and then collect more data with the objective of improving their predictive abilities. This leads to the well-known exploration (searching the space of possible decisions) and exploitation (choosing the optimal decision according to the learned model) dilemma. A principled way to capture this trade-off is the study of multi-armed bandit problems. On the other hand, these online learning systems are made up of several interacting components. Therefore, it is beneficial to study the pattern of interaction among these components, in order to explore in a sample efficient manner, which in turn leads to better exploitation. In this thesis, we will see that it is sometimes beneficial to view these online learning systems under the lens of causality; thus formalizing the pattern of interaction among the various components of the system, through causal graphical models. In our first problem, we study the contextual bandit problem with L observed contexts and K arms, with a latent low dimensional causal structure. We show that leveraging this latent low dimensional structure can lead to superior regret guarantees that are practical even for smaller time horizons. This also leads to the first regret guarantees for low-rank matrix completion where the rank is greater than one. Our second problem deals with leveraging information leakage in an online fashion in the presence of causal structures. We identify that in presence of general causal structures there is information leakage between different interventions viewed as arms of a bandit i.e. collecting data under one intervention can inform us about the statistics under other interventions. We demonstrate how to leverage this information leakage through adaptive importance sampling and apply our algorithm in biological networks and for interpretability of deep networks. This directly leads us to our third problem, where we use the idea of information leakage (explored in our second problem) in the context of stochastic contextual bandits. We propose the contextual bandits with stochastic experts problem and provide the first problem dependent regret bound in contextual bandits, where the scaling of the regret bound can potentially be as low as logarithmic in the number of experts. We show that our algorithm outperforms several state of the art algorithms on progressive validation tasks on multi-class classification data-sets. In our fourth problem, we look at conditional independence testing, which is one of the fundamental tools in causal structure learning. We reduce this problem into binary classification, through a nearest neighbor based bootstrap procedure. This enables us to use powerful supervised learning tools like gradient boosted trees or deep neural networks, that have desirable properties in higher dimensions. Finally in our… Advisors/Committee Members: Shakkottai, Sanjay (advisor), Caramanis, Constantine (committee member), Dimakis, Georgios-Alexandros (committee member), Johari, Ramesh (committee member), Sanghavi, Sujay (committee member).

Subjects/Keywords: Online learning; Multi-armed bandits; Contextual bandits; Hyper-parameter tuning; Tree-search; CI testing

10 more images…

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

-4677-643X. (2019). Online experiment design with causal structures. (Doctoral Dissertation). University of Texas – Austin. Retrieved from http://dx.doi.org/10.26153/tsw/2950

Note: this citation may be lacking information needed for this citation format:
Author name may be incomplete

Chicago Manual of Style (16^th Edition):

-4677-643X. “Online experiment design with causal structures.” 2019. Doctoral Dissertation, University of Texas – Austin. Accessed April 10, 2021. http://dx.doi.org/10.26153/tsw/2950.

Note: this citation may be lacking information needed for this citation format:
Author name may be incomplete

MLA Handbook (7^th Edition):

-4677-643X. “Online experiment design with causal structures.” 2019. Web. 10 Apr 2021.

Note: this citation may be lacking information needed for this citation format:
Author name may be incomplete

Vancouver:

-4677-643X. Online experiment design with causal structures. [Internet] [Doctoral dissertation]. University of Texas – Austin; 2019. [cited 2021 Apr 10]. Available from: http://dx.doi.org/10.26153/tsw/2950.

Note: this citation may be lacking information needed for this citation format:
Author name may be incomplete

Council of Science Editors:

-4677-643X. Online experiment design with causal structures. [Doctoral Dissertation]. University of Texas – Austin; 2019. Available from: http://dx.doi.org/10.26153/tsw/2950

Note: this citation may be lacking information needed for this citation format:
Author name may be incomplete

25. L. Cella. EFFICIENCY AND REALISM IN STOCHASTIC BANDITS.

Degree: 2021, Università degli Studi di Milano

URL: http://hdl.handle.net/2434/807862

► This manuscript is dedicated to the analysis of the application of stochastic bandits to the recommender systems domain. Here a learning agent sequentially recommends one… (more)

Subjects/Keywords: machine learning; multi-armed bandits; stochastic bandits; online learning; Settore INF/01 - Informatica

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Cella, L. (2021). EFFICIENCY AND REALISM IN STOCHASTIC BANDITS. (Thesis). Università degli Studi di Milano. Retrieved from http://hdl.handle.net/2434/807862

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Cella, L.. “EFFICIENCY AND REALISM IN STOCHASTIC BANDITS.” 2021. Thesis, Università degli Studi di Milano. Accessed April 10, 2021. http://hdl.handle.net/2434/807862.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Cella, L.. “EFFICIENCY AND REALISM IN STOCHASTIC BANDITS.” 2021. Web. 10 Apr 2021.

Vancouver:

Cella L. EFFICIENCY AND REALISM IN STOCHASTIC BANDITS. [Internet] [Thesis]. Università degli Studi di Milano; 2021. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/2434/807862.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Cella L. EFFICIENCY AND REALISM IN STOCHASTIC BANDITS. [Thesis]. Università degli Studi di Milano; 2021. Available from: http://hdl.handle.net/2434/807862

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Rochester Institute of Technology

26. AlOmar, Eman Abdullah. How We Refactor and How We Mine it ? A Large Scale Study on Refactoring Activities in Open Source Systems.

Degree: MS, Software Engineering (GCCIS), 2018, Rochester Institute of Technology

URL: https://scholarworks.rit.edu/theses/9784

► Refactoring, as coined by William Obdyke in 1992, is the art of optimizing the syntactic design of a software system without altering its external… (more)

▼ Refactoring, as coined by William Obdyke in 1992, is the art of optimizing the syntactic design of a software system without altering its external behavior. Refactoring was also cataloged by Martin Fowler as a response to the existence of design defects that negatively impact the software's design. Since then, the research in refactoring has been driven by improving systems structures. However, recent studies have been showing that developers may incorporate refactoring strategies in other development related activities that go beyond improving the design. In this context, we aim in better understanding the developer's perception of refactoring by mining and automatically classifying refactoring activities in 1,706 open source Java projects. We perform a it{differentiated replication} of the pioneering work by Tsantalis et al. We revisit five research questions presented in this previous empirical study and compare our results to their original work. The original study investigates various types of refactorings applied to different source types (i.e., production vs. test), the degree to which experienced developers contribute to refactoring efforts, the chronological collocation of refactoring with the release and testing periods, and the developer's intention behind specific types of refactorings. We reexamine the same questions but on a larger number of systems. To do this, our approach relies on mining refactoring instances executed throughout several releases of each project we studied. We also mined several properties related to these projects; namely their commits, contributors, issues, test files, etc. Our findings confirm some of the results of the previous study and we highlight some differences for discussion. We found that 1) feature addition and bug fixes are strong motivators for developers to refactor their code base, rather than the traditional design improvement motivation; 2) a variety of refactoring types are applied when refactoring both production and test code. 3) refactorings tend to be applied by experienced developers who have contributed a wide range of commits to the code. 4) there is a correlation between the type of refactoring activities taking place and whether the source code is undergoing a release or a test period. Advisors/Committee Members: Mohamed Wiem Mkaouer.

Subjects/Keywords: Commit classification; Mining software repository; Refractoring; Software evolution

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

AlOmar, E. A. (2018). How We Refactor and How We Mine it ? A Large Scale Study on Refactoring Activities in Open Source Systems. (Masters Thesis). Rochester Institute of Technology. Retrieved from https://scholarworks.rit.edu/theses/9784

Chicago Manual of Style (16^th Edition):

AlOmar, Eman Abdullah. “How We Refactor and How We Mine it ? A Large Scale Study on Refactoring Activities in Open Source Systems.” 2018. Masters Thesis, Rochester Institute of Technology. Accessed April 10, 2021. https://scholarworks.rit.edu/theses/9784.

MLA Handbook (7^th Edition):

AlOmar, Eman Abdullah. “How We Refactor and How We Mine it ? A Large Scale Study on Refactoring Activities in Open Source Systems.” 2018. Web. 10 Apr 2021.

Vancouver:

AlOmar EA. How We Refactor and How We Mine it ? A Large Scale Study on Refactoring Activities in Open Source Systems. [Internet] [Masters thesis]. Rochester Institute of Technology; 2018. [cited 2021 Apr 10]. Available from: https://scholarworks.rit.edu/theses/9784.

Council of Science Editors:

AlOmar EA. How We Refactor and How We Mine it ? A Large Scale Study on Refactoring Activities in Open Source Systems. [Masters Thesis]. Rochester Institute of Technology; 2018. Available from: https://scholarworks.rit.edu/theses/9784

27. Alawi A A. A Study of Commitment Schemes with Fuzziness incorporation;.

Degree: 2014, Symbiosis International University

URL: http://shodhganga.inflibnet.ac.in/handle/10603/31188

► The design of commitment schemes for securing many real world applications has recently gained significance. Commitment schemes are cryptographic protocols that ensure secure communication between… (more)

▼ The design of commitment schemes for securing many real world applications has recently gained significance. Commitment schemes are cryptographic protocols that ensure secure communication between two parties, with complete disillusionment of information for mistrusted parties. Such protocols are commonly implemented as two-phase processes (Commit and Open phases). In the Commit phase, one party (known as the Sender) seals a secret message in an envelope and sends the sealed envelope to the other party (known as the Receiver). In the Open phase, the Sender opens the sealed envelope which enables the Receiver to learn the original message. Protocols for commitment schemes have found applications ranging from cryptography to secure communication and biometric templates. newline newlineIn commitment schemes, commitments are conventionally opened using identical information. However, there could be many instances where the data transmission may be prone to channel noise. Also, in several important security applications the processed data is inherently noisy, such as biometric data. It is a well-known fact that this kind of uncertainty naturally leads to the introduction of a fuzziness dimension to the commitment problem at hand. Therefore, research in cryptography has aimed to design cryptographic primitives based on complexity assumptions that are as weak as possible, while it is commonly known that cryptographic primitives, like commitment schemes, based on hard complexity assumptions already exist. Then, it is intuitive to expect these existing commitment schemes to incorporate fuzziness. This thesis provides a thorough study and analysis of most of the existing commitment schemes that are based on variety complexity assumptions. The main findings of this study are: 1) first develop a general and mathematical formal framework for commitment schemes referred to as an ordinary crisp commitment scheme (OCCS). Most of existing conventional commitment schemes can be viewed as special types of OCCS schemes. The trusted third party is involved only in the setup phase of the scheme. As Electronic Commerce (E-Commerce) is becoming a day-to-day shopping style, it is desirable to have a mechanism that allows the two parties to sign a digital contract via the Internet in a fair way. Based on a variant of the OCCS scheme, a new digital contract signing protocol is proposed. The protocol has a simple functionality, and yet it satisfies the optimistic, fairness and abuse-free properties. Performance evaluation of the proposed protocol demonstrates its security and efficiency compared to other related protocols that are widely used in real-life applications. 2) Then, the fuzziness property is injected into the ordinary crisp commitment scheme to form a newline newlineXIV newline newlinenew and general mathematical framework for fuzzy commitment scheme referred to as the ordinary fuzzy commitment scheme (OFCS). Based on the proposed OFCS scheme, many fuzzy commitment schemes based on different complexity assumptions constructed. 3)… Advisors/Committee Members: Acharya Haridas S.

Subjects/Keywords: Commit and Open phases; Fuzziness incorporation , cryptographic protocols

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

A, A. A. (2014). A Study of Commitment Schemes with Fuzziness incorporation;. (Thesis). Symbiosis International University. Retrieved from http://shodhganga.inflibnet.ac.in/handle/10603/31188

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

A, Alawi A. “A Study of Commitment Schemes with Fuzziness incorporation;.” 2014. Thesis, Symbiosis International University. Accessed April 10, 2021. http://shodhganga.inflibnet.ac.in/handle/10603/31188.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

A, Alawi A. “A Study of Commitment Schemes with Fuzziness incorporation;.” 2014. Web. 10 Apr 2021.

Vancouver:

A AA. A Study of Commitment Schemes with Fuzziness incorporation;. [Internet] [Thesis]. Symbiosis International University; 2014. [cited 2021 Apr 10]. Available from: http://shodhganga.inflibnet.ac.in/handle/10603/31188.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

A AA. A Study of Commitment Schemes with Fuzziness incorporation;. [Thesis]. Symbiosis International University; 2014. Available from: http://shodhganga.inflibnet.ac.in/handle/10603/31188

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

University of Waterloo

28. Chong, Michael. Commit-Level vs. File-Level Vulnerability Prediction.

Degree: 2016, University of Waterloo

URL: http://hdl.handle.net/10012/10867

► Helping software development teams find and repair vulnerabilities before they are released and exploited can prevent costs due to loss of data, availability, and reputation.… (more)

Subjects/Keywords: Vulnerability Prediction; Commit-level Prediction; File-level Prediction

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Chong, M. (2016). Commit-Level vs. File-Level Vulnerability Prediction. (Thesis). University of Waterloo. Retrieved from http://hdl.handle.net/10012/10867

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Chong, Michael. “Commit-Level vs. File-Level Vulnerability Prediction.” 2016. Thesis, University of Waterloo. Accessed April 10, 2021. http://hdl.handle.net/10012/10867.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Chong, Michael. “Commit-Level vs. File-Level Vulnerability Prediction.” 2016. Web. 10 Apr 2021.

Vancouver:

Chong M. Commit-Level vs. File-Level Vulnerability Prediction. [Internet] [Thesis]. University of Waterloo; 2016. [cited 2021 Apr 10]. Available from: http://hdl.handle.net/10012/10867.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Chong M. Commit-Level vs. File-Level Vulnerability Prediction. [Thesis]. University of Waterloo; 2016. Available from: http://hdl.handle.net/10012/10867

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

29. Alan Keller Gomes. Representação, extração e avaliação de interações entre usuários de redes sociais online.

Degree: 2013, University of São Paulo

URL: http://www.teses.usp.br/teses/disponiveis/55/55134/tde-26082013-105631/

►

Com a popularidade e o crescimento das Redes Sociais Online, o interesse pelo entendimento de como seus usuários interagem entre si também tem crescido, configurando… (more)

▼

Com a popularidade e o crescimento das Redes Sociais Online, o interesse pelo entendimento de como seus usuários interagem entre si também tem crescido, configurando assim um cenário rico no qual são criadas oportunidades para melhorar o design de interfaces, compreender a organização de movimentos sociais, o marketing viral e a distribuição de conteúdos multimídia, dentre outros. Na investigação dessas oportunidades de pesquisa, um modelo de representação da atividade dos usuários amplamente empregado é baseado na construção de um grafo. Embora as atividades dos usuários em uma Rede Social Online sejam variadas, assim como as possibilidades de interação entre usuários, a construção de um grafo normalmente considera uma interação específica, que é então analisada a partir da interpretação de medidas baseadas em grafos ou de medidas estatísticas. Técnicas de mineração de dados podem ser empregadas de forma alternativa e complementar ao modelo baseado em grafos, possibilitando a extração e a avaliação de padrões das atividades de usuários. Entretanto, tanto na análise baseada em grafos quanto na utilização de técnicas de mineração de dados, a literatura reporta trabalhos em que são implicitamente representados elementos associados com a atividade dos usuários como ações executadas, mídias compartilhadas, aplicações e tipos de dispositivos utilizados. Nesse cenário, existe uma demanda por um modelo descritivo que permita a representação explícita dos elementos associados com as atividades dos usuários representação essa que possa ser utilizada na extração e na avaliação das interações entre usuários. Para atender essa demanda, nesta tese é apresentada uma técnica e um método para representar, extrair e avaliar interações entre usuários de Redes Sociais Online. São também reportados resultados de experimentos da aplicação da técnica e do método a partir de dados obtidos de uma Rede Social Online

The interest in understanding user interactions is increasing with the popularity of Online Social Networks. This scenario leads to rich opportunities to improve the design of interfaces, to understand the organization of social movements, to model the dynamics of viral marketing and to characterize the distribution of multimedia content, among others. In the investigation of these research opportunities, a widely used representation of users activities is based on graphs as underlying data models. Graph data models usually consider a specific interaction, which is analyzed based on the interpretation of statistical measures and graph-based measures. On the other hand, user activities and possibilities of interaction in social networks are multifaceted, so that singleinteraction graph-based approaches might prove to be unfeasible to properly model the problem. In order to deal with these limitations, data mining techniques may be employed as an alternative to graph-based modeling approaches, by enabling extraction and evaluation of users activities as patterns. However, both in the use of the model graph-based as in…

Advisors/Committee Members: Maria da Graça Campos Pimentel, Renata Pontin de Mattos Fortes, Jose Valdeni de Lima, Ethan Vincent Munson, Vânia Paula de Almeida Neris.

Subjects/Keywords: Contigências comportamentais; Interações sociais; Redes sociais online; Regras if-then; Behavioral contingencies; If-then rules; Online social networks; Social interactions

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Gomes, A. K. (2013). Representação, extração e avaliação de interações entre usuários de redes sociais online. (Doctoral Dissertation). University of São Paulo. Retrieved from http://www.teses.usp.br/teses/disponiveis/55/55134/tde-26082013-105631/

Chicago Manual of Style (16^th Edition):

Gomes, Alan Keller. “Representação, extração e avaliação de interações entre usuários de redes sociais online.” 2013. Doctoral Dissertation, University of São Paulo. Accessed April 10, 2021. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-26082013-105631/.

MLA Handbook (7^th Edition):

Gomes, Alan Keller. “Representação, extração e avaliação de interações entre usuários de redes sociais online.” 2013. Web. 10 Apr 2021.

Vancouver:

Gomes AK. Representação, extração e avaliação de interações entre usuários de redes sociais online. [Internet] [Doctoral dissertation]. University of São Paulo; 2013. [cited 2021 Apr 10]. Available from: http://www.teses.usp.br/teses/disponiveis/55/55134/tde-26082013-105631/.

Council of Science Editors:

Gomes AK. Representação, extração e avaliação de interações entre usuários de redes sociais online. [Doctoral Dissertation]. University of São Paulo; 2013. Available from: http://www.teses.usp.br/teses/disponiveis/55/55134/tde-26082013-105631/

UCLA

30. Buffington, Joseph Patrick. The Logical Form of Contract Formation.

Degree: Linguistics, 2015, UCLA

URL: http://www.escholarship.org/uc/item/72k468x2

► This dissertation lays the foundation for linguistic inquiry into the question of how speakers of English know when a contractual offer has been made. Of… (more)

Subjects/Keywords: Linguistics; Law; conditionals; contracts; formation; if-then; offers; promises

Record Details Similar Records Cite « Share

❌

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

APA (6^th Edition):

Buffington, J. P. (2015). The Logical Form of Contract Formation. (Thesis). UCLA. Retrieved from http://www.escholarship.org/uc/item/72k468x2

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Chicago Manual of Style (16^th Edition):

Buffington, Joseph Patrick. “The Logical Form of Contract Formation.” 2015. Thesis, UCLA. Accessed April 10, 2021. http://www.escholarship.org/uc/item/72k468x2.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

MLA Handbook (7^th Edition):

Buffington, Joseph Patrick. “The Logical Form of Contract Formation.” 2015. Web. 10 Apr 2021.

Vancouver:

Buffington JP. The Logical Form of Contract Formation. [Internet] [Thesis]. UCLA; 2015. [cited 2021 Apr 10]. Available from: http://www.escholarship.org/uc/item/72k468x2.

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

Council of Science Editors:

Buffington JP. The Logical Form of Contract Formation. [Thesis]. UCLA; 2015. Available from: http://www.escholarship.org/uc/item/72k468x2

Note: this citation may be lacking information needed for this citation format:
Not specified: Masters Thesis or Doctoral Dissertation

		in
And / Or
		in
And / Or
		in
And / Or
		in

Open Access Theses and Dissertations