Artificial Intelligence 15-780 and 16-731

Welcome! This class is in Wean 5409 (though for some lectures, students taking 16-731 will meet in Smith Hall 101), Tuesdays and Thursday 10:30-11:50, starting Tuesday Jan 13th. The course book is Russell and Norvig's "Artificial Intelligence: A Modern Approach". The instructors are

Course Description

The graduate AI Core course focuses on fundamental techniques for automated problem solving: symbolic, numeric and hybrid. These techniques include: state-space search, constraint propagation, logic-based and frame-based representations, elements of game theory, elements of decision theory and optimization, elements of information theory, induction of decision trees, reinforcement learning, Bayesian networks, nearest-neighbor/case-based/analogical methods, and so on. The kinds of reasoning tasks addressed via these techniques include classification-based decision-making, planning, natural-language processing, speech recognition, and several robotic tasks.

The Robotics and Engineering variant of the course (16-731) will cover some of the engineering, industrial, and robotics-specific topics in more detail. These extra topics include robot motion planning, linear programming, learning to control dynamic systems, industrial scheduling. 16-731 also covers the topics in state-space search differently, spending more time on the lower level data structures and algorithms and less time on the higher level issues of abstraction.

Here is the usual list of administrative course details:

  • Prerequisites: Programming experience, mathematical background (preferably some knowledge of statistics). Some knowledge of algorithms, of AI (e.g. 15-381) and of robotics is preferred.

  • Grading: Approximately four homework assignments (together counting for approximately 30% of final grade), midterm (20%), final exam (30%) and project (approximately 20%).

  • Workload: Moderate. 12 hours/week estimate.

  • Assignments (anticipated): 2 problem sets, 2 hands-on programming tasks.

  • Projects: To be performed by teams of 2 or 3 students (all will receive the same grade in the project). Project topics will be distributed later. The homeworks will be given mostly in the first half of the course to permit work on the projects in the second half.

    Syllabus
    Day Subject Lecturer Assignment
    Knowledge Representation
    1/13 Course introduction. Logic and predicate calculus. Andrew
    1/15 Unification, Resolution for theorem proving. Andrew
    1/20 Representing uncertain knowledge: Probability refresher, Bayesian inference, Belief Network knowledge representation. Andrew
    1/22 Inference in Belief Networks. Andrew
    Search (15-780 CS variant)
    1/27 Heuristic Search Manuela
    1/29 State-space, plan-space planning Manuela
    2/3 Hierarchical planning Manuela
    2/5 Other algorithms; Comparison Manuela
    2/10 Probabilistic planning Manuela
    2/12 Plan execution in uncertain domains Manuela
    Search (16-731 robotics and engineering variant)
    1/27 Search Algorithms Andrew
    1/29 Advanced Data Structures for Search Algorithms Andrew
    2/3 A-star search Andrew
    2/5 Robot Motion Planning Andrew
    2/10 Multi-path planning Andrew
    2/12 The D* search method for real-time vehicle control Tony Stentz
    Applied Search
    2/17 Constraint satisfaction and scheduling Andrew
    2/19 Stochastic Optimization, Sim. annealing, GA's Manuela
    2/24 Game playing, game theory Andrew
    2/26 TBD TBD
    3/3 Midterm Exam, open book, WeH 5409
    Machine Learning
    3/5 Decision Trees and Version Spaces Tom
    3/10 Neural networks Tom
    3/12 Learning Bayes Nets Tom
    3/17 EM and Clustering Tom
    3/19 TBD TBD
    3/24 Spring Break
    3/26 Spring Break
    3/31 Cognitive Science Simon
    4/2 TBD TBD
    Nondeterministic Action Selection and Learning
    4/9 Markov Decision Processes Andrew
    4/14 Reinforcement Learning Manuela
    4/16 Hidden Markov Models Andrew
    4/21 Multiagent systems Manuela
    Advanced Topics (CS 15-780 variant)
    4/23 Relational learning Tom
    4/28 Learning and planning Manuela
    4/30 Perception: NLP and vision Tom
    Advanced Topics (Robotics and Engineering 16-731 variant)
    4/23 Optimization with Linear Programming Andrew
    4/28 Learning Control of Dynamic Systems Andrew
    4/30 Control and Optimization of large-scale hierarchical systems Andrew
    5/5 Final Exam, WeH 5409

    Last modified 11/25/97
    Tom Mitchell ( tom.mitchell@cs.cmu.edu)