Project Outline

Students should form teams of 2-3 (preferably 3) per project. Projects can be of the following types Article-centered project. For details see below. Programming project. For details see below. Your own project idea, on the topic of AI (e.g., implementing an algorithm and applying to some domain). Students need to obtain approval before proceeding. Each team needs to submit a written report (one report per team) at the end of the course (exact date TDB). Teams should be formed and project proposals finalized/approved by early Nov at the latest. Article-centered project A project can be centered on 1-2 primary articles selected from a recent AI conference/journal. A list of selected conferences is below, as well as a list of selected papers. The idea is that the paper will open a window to a subject that the students read and explore. You may need to read earlier papers for background and information (that appear in the citations of the primary papers). Students may investigate the validity of paper by running code developed on some benchmarks whenever relevant. Students can communicate with authors to get slides for presentations and code. The team will present the paper to the class and write a project report to be handed in at the end of the course. The project report should include a description and evaluation of the paper and any creative independent work carried out by the team. Programming project A project can consist of picking a particular problem class and task (e.g. solving Constraint Satisfaction Problems), and writing a computer program to solve problem instances of this class. There are repositories of benchmarks that students can get problem instances, for example UCI repository Cost Function Optimization problems Selected Conferences CP 2012 Website Proceedings CP 2011 Website Proceedings SoCS 2013 Website Proceedings SoCS 2012 Website Proceedings SoCS 2011 Website Proceedings ICAPS 2013 Website Proceedings ICAPS 2012 Website Proceedings ICAPS 2011 Website Proceedings ICAPS 2008 Website Proceedings ICAPS 2007 Website Proceedings Selected Papers Below is a list selected papers from various conferences. Feel free to consider the full lists of conferences listed above. A Survey of Monte Carlo Tree Search Methods (in IEEE Transactions On Computational Intelligence and AI in Games, 2012) Cameron Browne, et. al. Width and Serialization of Classical Planning Problems (in ECAI 2012) Nir Lipovetzky and Hector Geffner Width and Complexity of Belief Tracking in Non-Deterministic Conformant and Contingent Planning (in AAAI 2012) Blai Bonet and Hector Geffner Action Selection for MDPs: Anytime AO* vs. UCT (in AAAI 2012) Blai Bonet and Hector Geffner Planning as Satisfiability: Heuristics (in AIJ 2012) Jussi Rintanen SoCS 2012 Are We There Yet? - Estimating Search Progress Jordan T. Thayer, Roni Stern, and Levi H. S. Lelis Real-Time Motion Planning with Dynamic Obstacles Jarad Cannon, Kevin Rose, and Wheeler Ruml Execution Ordering in AND/OR Graphs with Failure Probabilities Priyankar Ghosh, P. P. Chakrabarti, and Pallab Dasgupta Performance Analysis of Planning Portfolios Sergio Núñez, Daniel Borrajo, and Carlos Linares López Multimapping Abstractions and Hierarchical Heuristic Search Bo Pang and Robert C. Holte When does Weighted A* Fail? Christopher Wilt and Wheeler Ruml SoCS 2011 Cost-Based Heuristic Search Is Sensitive to the Ratio of Operator Costs Christopher Makoto Wilt, Wheeler Ruml Adapting a Rapidly-Exploring Random Tree for Automated Planning Vidal Alcázar, Manuela Veloso, Daniel Borrajo Best-First Search for Bounded-Depth Trees Kevin Rose, Ethan Burns, Wheeler Ruml All PSPACE-Complete Planning Problems Are Equal but Some Are More Equal than Others Christer Backstrom, Peter Jonsson Degrees of Separation in Social Networks Reza Bakhshandeh, Mehdi Samadi, Zohreh Azimifar, Jonathan Schaeffer Automatic Move Pruning in General Single-Player Games Neil Burch, Robert C. Holte Deadline-Aware Search Using On-Line Measures of Behavior Austin J. Dionne, Jordan Tyler Thayer, Wheeler Ruml These two papers can be a single project for a team of 3. Path Planning with Adaptive Dimensionality Kalin Gochev, Benjamin Cohen, Jonathan Butzke, Alla Safonova, Maxim Likhachev Search-Based Planning with Provable Suboptimality Bounds for Continuous State Spaces Juan Pablo Gonzalez, Maxim Likhachev A Novel Technique for Compressing Pattern Databases in the Pancake Sorting Problems Morteza Keshtkaran, Roohollah Taghizadeh, Koorush Ziarati A Preliminary Evaluation of Machine Learning in Algorithm Selection for Search Problems Lars Kotthoff, Ian P. Gent, Ian Miguel Faster Optimal and Suboptimal Hierarchical Search Michael J. Leighton, Wheeler Ruml, Robert C. Holte These two papers are a single project. Predicting Solution Cost with Conditional Probabilities Levi Lelis, Roni Stern, Shahab Jabbari Arfaee Improved Prediction of IDA*'s Performance via Epsilon-Truncation Levi Lelis, Sandra Zilles, Robert Craig Holte State-Set Search Bo Pang, Robert C. Holte Size-Independent Additive Pattern Databases for the Pancake Problem Álvaro Torralba Arias de Reyna, Carlos Linares López Scalable Distributed Monte-Carlo Tree Search Kazuki Yoshizoe, Akihiro Kishimoto, Tomoyuki Kaneko, Haruhiro Yoshimoto, Yutaka Ishikawa ICAPS 2012 Tractable Monotone Temporal Planning [PDF] Martin C. Cooper, Frederic Maris, Pierre Regnier Sampling-Based Coverage Path Planning for Inspection of Complex Structures [PDF] Brendan J. Englot, Franz S. Hover Pruning Methods for Optimal Delete-Free Planning [PDF] Avitan Gefen, Ronen Brafman Incremental Lower Bounds for Additive Cost Planning Problems [PDF] Patrik Haslum Optimal Search with Inadmissible Heuristics [PDF] Erez Karpas, Carmel Domshlak PROST: Probabilistic Planning Based on UCT [PDF] Thomas Keller, Patrick Eyerich Semi-Relaxed Plan Heuristics [PDF] Emil Ragip Keyder, Joerg Hoffmann, Patrik Haslum Reverse Iterative Deepening for Finite-Horizon MDPs with Large Branching Factors [PDF] Andrey Kolobov, Peng Dai, Mausam Mausam, Daniel S. Weld Predicting Optimal Solution Cost with Bidirectional Stratified Sampling [PDF] Levi Lelis, Roni Stern, Ariel Felner, Sandra Zilles, Robert C. Holte On Computing Conformant Plans Using Classical Planners: A Generate-And-Complete Approach [PDF] Khoi Hoang Nguyen, Vien Dang Tran, Tran Cao Son, Enrico Pontelli Incremental ARA*: An Incremental Anytime Search Algorithm for Moving-Target Search [PDF] Xiaoxun Sun, William Yeoh, Tansel Uras, Sven Koenig Faster Bounded-Cost Search Using Inadmissible Estimates [PDF] Jordan Tyler Thayer, Roni Stern, Ariel Felner, Wheeler Ruml Short-Sighted Stochastic Shortest Path Problems [PDF] Felipe W. Trevizan, Manuela M. Veloso Bandit-Based Planning and Learning in Continuous-Action Markov Decision Processes [PDF] Ari Weinstein, Michael L. Littman Planning Via Random Walk-Driven Local Search [PDF] Fan Xie, Hootan Nakhost, Martin Müller