The program is held in three terms, starting September 2014. The courses address learning, action, and interaction in real or virtual environments, that may include humans or other agents.
- Machine Learning (Anders Jonsson)
- Autonomous Systems (Hector Geffner)
- Mobile Robotics (Vladimir Estivill-Castro)
- Research Methodology (Davinia Hernández-Leo)
- Elective 2
- Master Thesis
The two electives from a list that includes:
- Pattern Recognition
- Computer Vision
- Virtual Communication Environments
- Face and Gesture Analysis
- Cloud Computing
- Signal Processing
- Speech Technologies
Not all electives taught every year though, and offer conditional upon sufficient student enrollment.
Machine learning has achieved a great importance in recent years due to amount of data that is being collected that cannot be efficiently processed by humans. Learning sytems are used in a number of applications including recommending systems, spam filtering, etc. The course covers a number of machine learning formulations and algorithms: from supervised methods, where information provided by a teacher in the form of samples needs to be generalized to unseen situations, to unsupervised methods that learn from experience. The former methods include the induction of decision trees from data, the perceptron algorithm, and support vector machines, while the latter include reinforcement learning. We also cover statistical learning methods and the theoretical aspects underlying learning approaches, so-called Computational Learning Theory.
- Computational Learning Theory
- Linear Discriminant Functions: Perceptrons, Support Vector Machines
- Non-Metric Methods: Learning Decision Trees
- Clustering: K-means and Mixture Models
- Reinforcement Learning
- T. Mitchell: Machine Learning, McGraw Hill, 1997
- C. M. Bishop, Pattern Recognition and Machine Learning, Springer, 2007
- R. Sutton and A. Barto, Reinforcement Learning: An Introduction, MIT Press, 1998
- Y. Abu-Mostafa, M. Magdon-Ismail & H-T. Lin: Learning from Data. AMLBook, 2012
- Exam 50 %
- Projects: 50 %
- Theory: Professor Lectures
- Project Homework [top]
The focus of this course is autonomous behavior, and more precisely, the different methods for developing "agents" capable of making their own decisions in real or simulated environments. This includes characters in video-games, robots, softbots in the web, etc. The problem of developing autonomous agents is a fundamental problem in Artificial Intelligence, where three basic approaches have been developed: the programmer-based approach, where the agent responses are hardwired by a human programmer; the learning-based approach, where the agent learns to control its behavior from experience or information obtained from a teacher, and the model-based approach, where the agent control is derived automatically from a model describing the goals, the actions available, and the sensing capabilities. In the course, we review the three approaches to developing autonomous systems, with emphasis on the model-based approach, which in AI goes under the name of planning. We study autonomy in dynamic, partially observable settings involving a single agent or multiple agents. The course involves theory and experimentation.
- Autonomous Behavior: Introduction, Approaches
- Model 1: Classical Planning Models
- Model 2: Markov Decision Processes
- Model 3: Partially Observable MDPs
- Model-based Solution methods: Heuristic Search, Value and Policy Iteration, LRTA*, RTDP
- Model-free solution methods: Reinforcement Learning
- Programming the control: finite-state machines and controllers
- Multi-agent Models: fundamentals, Game Theory, Multi-agent planningIntroduction
- Artificial Intelligence: A Modern Approach, S. Russell and P. Norvig, Prentice Hall (3rd Edition), 2009
- Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations, Yoav Shoham and Kevin Leyton-Brown, Cambridge Univ. Press, 2008
- Automated Planning: Theory & Practice, Malik Ghallab, Dana Nau, Paolo Traverso, Morgan Kaufmann, 2004
- Reinforcement Learning: An Introduction by Richard S. Sutton and Andrew G. Barto, MIT Press 1998
- Neuro-Dynamic Programming, Dimitri P. Bertsekas and John N. Tsitsiklis, Athena Scientific 1996
- Programming Game AI by Example, Mat Buckland, Jones & Bartlett Publishers; 2004
- Exam 50 %
- Programming Projects: 25 %
- Seminar Presentations 25 %
- 20h Theory: Professor Lectures
- 5h Student Seminars
- Project Homework [top]
Introduction to mobile robotics covering practical and theoretical aspects. Course will involve basic notions of robot locomotion, perception, localization, and action; robot architectures, and projects on real robots.
- Introduction: problem statements, typical applications.
- Locomotion with legs and wheels
- Mobile Robots Kinematics and motion control
- Perception 1: Sensors
- Perception 2: Uncertainties, Line Features
- Perception 3: Vision
- Localization 1: Introduction and odometry error model
- Localization 2: Probabilistic map based localizaation and Markov localization
- Localization 3: Kalman filter localization
- Planning and Navigation: Introduction, path planning, obstacle avoidance, decomposition
- Agent architectures for mobile robots
- Introduction to autonomous mobile robots (2nd), Roland Siegwart and Illah R. Nourbakhsh, MIT Press, 2011.
- Mobile robots : inspiration to implementation (2nd) ,Joseph L. Jones, Anita M. Flynn,A K Peters, 1999 Edició 2nd. ed.
- Autonomous robots : from biological inspiration to implementation and control / George A. Bekey, MIT Press, 2005
- Probabilistic robotics, S. Thrun, Wolfram Burgard, Dieter Fox, MIT Press, 2005
- Projects 40%
- Exams 30%
- Oral Presentations 30% [top]
The couse covers the central themes involved in the interaction with intelligent agents through the use of natural language, with emphasis on dialogue and language generation. We will also study planning techniques applied to the theory of speech acts and the use of rhetorical structures, both for
controlled dialogues as for dynamic and non-cooperative dialogues. Regarding analysis and generation of language, students will learn robust and incremental techinques capable of dealing with partial, and even ungrammatical discourse, as it's typical of spontaneous dialogues. We will also look at the design of dialogue architectures, and analyze the use of dialogue in "chatbots" and videogames.
- Introduction: History of human-computer interaction, problem statements, typical applications.
- Dialogue models, basic types of dialogues (conversational analysis, principles and characteristics of cooperative and non-cooperative dialogues), pragmatics.
- Dialogue strategies (types of dialogue strategies, relation to the game theory); speech act and discourse structure driven dialogue strategies.
- User modeling in dialogue systems.
- Dialogue languages.
- Machine learning techniques in dialogue systems.
- Linguistic background for natural language parsing (analysis) and generation.
- Parsing 1: Basic syntactic parsing techniques
- Parsing 2: Robust semantic parsing
- Generation 1: Stochastic generation techniques
- Generation 2: Incremental language generation
- Design of dialogue systems
- Applications 1: Case study of dialogues in selected chatbots
- Applications 2: Case study of dialogues in selected role-oriented video games.
- Project 50%
- Exam 30%
- Presentations 20%
- ... [top]
Study how to gather, process, search and mine data in the Web and its applications to search engines. Understand the basic concepts behind information retrieval and data mining.
The student must have a solid background in mathematics, algorithms and data structures. Machine learning is optional but suggested.
Its content is divided in three parts: theory of information retrieval, information retrieval on the web and information retrieval on data.
- Characteristics of the Web. Web structure. Retrieval vs. browsing. The long tail. Social networks.
- Basic concepts of information retrieval and data mining
- Main document relevance models: Boolean, vector, probabilistic. Browsing models. Precision vs. retrieval. Quality evaluation. Reference collections.
- Inverted indexes. Construction. Query processing. Use of compression.
- Basic data mining algorithms.
- Information retrieval on the web
- Architecture of a Web search engine. The crawler. Indexing systems, queries and ranking. Scalability. Ranking through link analysis. Multimedia search: images, audio and video.
- Web Data Mining
- Mining the content of the Web. Example: opinion mining. Structure Mining and Social Networks. Example: finding communities. Usage mining. Example: query log analysis. Advanced example: Web Spam detection.
- Baeza-Yates, R. y Ribeiro-Neto, B. Modern Information Retrieval, 2nd edition, Addison-Wesley 2010 (www.mir2ed.org)
- Chakrabarti,. S. Web data mining: Discovering Knowledge from Hypertext Data, Morgan Kaufmann, 2002 (second edition to appear).
- World Wide Web Consortium, w3c.org, 2011.
- The evaluation is through a written report where the student has to survey a topic related to the content of the course, including his/her own thoughts. [top]
- Research Project (20 ECTS): Carry out a research project and write a thesis under the supervision of a teacher. Includes a weekly class to present and discusss relevant topics to help decide, develop and present the thesis work. [top]