Skip to content. | Skip to navigation

Personal tools
Log in
Sections
You are here: Home People Kee-Eung Kim courses 2008 cs788

CS788 Dialogue Management Systems

by Kee-Eung Kim last modified Jun 21, 2011 02:33 PM

Overview

We will study computational models and algorithms for the machine to "intelligently" interact with the user. We will particularly focus on decision theoretic view of interaction. Hence, if you are looking for a course on HCI design methodologies and principles, you are looking at a wrong course.

First half of the semester will be run by instructor's lectures. Second half will be run by student presentations.

Student presentations: Each student will have to make 40-minute presentation for at least one paper. Those who are not presenting should read the assigned paper(s) and submit a one page write-up containing

  • summary of the paper in English (at least 700 words)
  • at least 3 research questions / ideas (in other words, don't just read the paper; think!)

at the beginning of each class. Upload your homework to the course bbs.

Instructor

Kee-Eung Kim

Prerequisites

Undergrad/grad artificial intelligence or equivalent (i.e., if you don't have basic knowledge about artificial intelligence techniques, you won't be able to catch the concepts)

Grading Policy

Midterm exam: 30%
Homework assignments: 40%
Presentation: 30%

If you don't submit at least one homework, I cannot guarantee anything about your final grade. If you have a business travel and have to miss the class, the homework is still due on the scheduled date. Late homeworks are not accepted.

Schedule

9/1 Orientation & Logistics [slides]
9/3 Dialogue Systems McTear, Spoken Dialogue Technology: Enabling the Conversational User Interface, ACM Computing Surveys 34 (1), 2002 
[slides]
9/5 Dialogue Systems
9/8 Markov Decision Processes (MDPs) [slides]
9/10 Reinforcement Learning [slides]
Singh, Litman, Kearns, and Walker, Optimizing Dialogue Management with Reinforcement Learning: Experiments with the NJFun System, Journal of Artificial Intelligence Research, 2002
9/17 Reinforcement Learning
9/22-9/26 No class I'll be at Interspeech 2008
9/29 POMDPs [slides]
10/1 POMDPs
10/6
10/8
(PO)MDPs for Dialogue Management Schatzmann, Stuttle, Weilhammer, and Young, Effects of the User Model on Simulation-based Learning of Dialogue Strategies, IEEE Automatic Speech Recognition and Understanding Workshop 2005
Williams & Young, Partially Observable Markov Decision Process for Spoken Dialog Systems, Computer Speech and Language 21, 2007
10/13 POMDP algorithms Kaelbling, Littman, and Cassandra, Planning and Acting in Partially Observable Stochastic Domains, Artificial Intelligence 101, 1998
10/15 POMDP algorithms Roy, Pineau, and Thrun, Spoken Dialogue Management Using Probabilistic Reasoning, Proceedings of ACL, 2000
10/20-10/24 Midterm exam week
10/27 POMDP algorithms
최재득
Thrun, Monte Carlo POMDPs, Proceedings of NIPS, 2000
10/29 POMDP algorithms
강병곤
Hansen, Solving POMDPs by Searching in Policy Space, Proceedings of UAI, 1998
11/3 POMDP algorithms 
이재승
Poupart & Boutilier, Bounded Finite State Controllers, Proceedings of NIPS, 2004
11/5 POMDP algorithms
석재현
Pineau, Gordon, and Thrun, Anytime point-based approximations for large POMDPs, Journal of Artificial Intelligence Research, 2006
11/10 POMDP algorithms
정제현
Ji, Parr, Li, Liao and Carin, Point-Based Policy Iteration, Proceedings of AAAI, 2007
11/12

POMDP algorithms
김영욱

Williams & Young, Scaling POMDPs for Spoken Dialog Management, IEEE Transactions on Audio, Speech and Language Processing, 2007
11/17 POMDP algorithms
배수정
Smith & Simmons, Heuristic Search Value Iteration for POMDPs, Proceedings of UAI, 2004
Smith & Simmons, Point-based POMDP algorithms: Improved analysis and implementation, Proceedings of UAI, 2005
11/19 POMDP algorithms
김진석
Sim, Kim, Kim, Chang, and Koo, Symbolic HSVI for Factored POMDPs, Proceedings of AAAI, 2008
11/24 POMDP algorithms
최재득
Shani, Brafman, Shimony, and Poupart, Efficient ADD Operations for Point-Based Algorithms. Proceedings of ICAPS, 2008
11/26 POMDP algorithms
이재승
Shani, Brafman, and Shimony, Forward Search Value Iteration for POMDPs, Proceedings of IJCAI, 2007
12/1 POMDP algorithms
정제현
Kurniawati, Hsu, and Lee, SARSOP: Efficient Point-Based POMDP Planning by Approximating Optimally Reachable Belief Spaces, Proceedings of Robotics: Science and Systems, 2008
12/3 POMDP algorithms
배수정
Hoey & Poupart, Solving POMDPs with Continuous or Large Discrete Observations Spaces, Proceedings of IJCAI, 2005
12/8 POMDP algorithms
석재현
Williams, The Best of Both Worlds: Unifying Conventional Dialog Systems and POMDPs, Proceedings of Interspeech, 2008
12/10 POMDP algorithms
김진석
Kim, Sim, Kim, Kim, Kim, and Sung, Effects of User Modeling on POMDP-Based Dialogue Systems, Proceedings of Interspeech, 2008
12/15-12/19 Final exam week