CPSC 522 - Artificial Intelligence 2
Schedule

Here is a tentative schedule. Topics in the future should be regarded as science fiction. Which students are presenting on each day: http://wiki.ubc.ca/Course:CPSC522/StudentPresentations2020.

See also videos.

• Jan 6 - AI and Agents. Slides: lect1.pdf, lect2.pdf. agents, AIpython (agent.py was demoed). Readings: AIFCA 2e: chapter 1. Sections 2.1-2.2.
• Jan 8 - Agents and Beliefs hierarchical control, probability, independence and belief networks. independence consequences. Readings: AIFCA 2e: Section 2.1-2.2, 8.1-8.3. Belief and Decision networks applet (go to downloads).
• Jan 13 - Probability & Graphical Models. independence, Markov networks, representing CPDs , inference. Readings: AIFCA 2e: Sections 8.3-8.4. Discussion papers (everyone):
  • Artificial Intelligence - The Revolution Hasn't Happened Yet,
  • Human-level intelligence or animal-like abilities? (and a video at ucla). (These should both be free downloads from UBC).
• Jan 15 - Probabilistic inference (cont). Classes cancelled due to snow. inference slides, Video part 1. Video part 2. AISpace downloads (download the belief and decision networks app). Readings: AIFCA 2e: Section 8.4.
• Jan 20 - Markov models. Localization demo: Localization demo (you might need to download Java code). Exact Inference - Discussion
  • Tian Sang, Paul Beame, and Henry A. Kautz. Performing Bayesian inference by weighted model counting. In AAAI, 2005.
• Jan 22 - Stochastic Simulation, Learning Probabilites. AIspace belief network example: http://artint.info/code/aispace/beta.xml.
• Jan 27 - Learning Probabilites, K-means and EM. Discussion paper:
  • Monte Carlo Localization: Efficient Position Estimation for Mobile Robots. D. Fox, Burgard, W., Dellaert, F., Thrun, S., AAAI, 1999
• Jan 29 - Learning (cont) Learning Belief Networks. See also Information Theory. AIspace belief network examples: http://artint.info/code/aispace/topic.xml, http://artint.info/code/aispace/sprinklerseason.xml.
• Feb 3 - Causality. AIspace example: http://artint.info/tutorials/causality/marijuana.xml. Discussion
  • Martin Trapp, Robert Peharz, Hong Ge, Franz Pernkopf, Zoubin Ghahramani, Bayesian Learning of Sum-Product Networks, Proc Neurips 2019.
• Feb 5 - Probabilistic Relational Models
• Feb 10 - Class cancelled Was: Discussion paper:
  • Shpitser, I., K. Mohan, and J. Pearl (2015). Missing data as a causal and probabilistic problem. In Proceedings of the Thirty-First Conference on Uncertainty in Artificial Intelligence
• Feb 12 - Probabilistic Relational Models (cont)., Relational probabilistic models.
• Feb 24 - Relational and embedding-based models
• Feb 26 - Lifted Inference, preferences and utility. Textbook. Was: Discussion paper:
  • Trouillon, T.; Welbl, J.; Riedel, S.; Gaussier, E.; and Bouchard, G. 2016. Complex embeddings for simple link prediction. In ICML, 2071-2080.
• Mar 2 - Preferences and Utility (c0nt) decision networks.
• Mar 4 - decision networks see Section 9.2 of textbook. Discussion paper:
  • Simultaneous Elicitation of Preference Features and Utility. Craig Boutilier, Kevin Regan and Paolo Viappiani. Proceedings of the Twenty-fourth AAAI Conference on Artificial Intelligence (AAAI-10) , pp.1160--1167, Atlanta GA (2010).
• Mar 9 - decision processes. see Section 9.2 of textbook. Value Iteration Applet.
• Mar 11 - reinforcement learning. See AIPython reinforcement learning code or Chapter 12 code. Discussion paper:
  • Computational rationality: A converging paradigm for intelligence in brains, minds, and machines, Samuel J. Gershman, Eric J. Horvitz, Joshua B. Tenenbaum, Science Vol. 349, Issue 6245, pp. 273-278
• Mar 16 - Reinforcement Learning. slides. Video part 1, Video part 2.
• Mar 18 - RL Part 3. RL Part 4 . Discussion paper:
  • Say, B., Wu, G., Zhou, Y.Q., Sanner, S.: Nonlinear hybrid planning with deep net learned transition models and mixed-integer linear programming. In: Twenty- Sixth International Joint Conference on Artificial Intelligence. pp. 750-756 (2017).
  • Obada's presentation
• Mar 23 -Reinforcement Learinng Part 5.
• Mar 25 - Discussion paper:
  • Kocsis, L., and Szepesvari, C. 2006. Bandit Based Monte-Carlo Planning. In Proceedings of the 17th European Conference on Machine Learning (ECML), 282-293.
• Mar 30, Apr 1 - Multiagent systems Lecture 1 Lecture 2 Lecture 3.
• Apr 1 - Discussion paper:
  • David Silver at al. Mastering the game of go without human knowledge. Nature, 550:354-359, 2017. (see also Deep Reinforcement Learning Doesn't Work Yet and Towards Generalization and Simplicity in Continuous Control) .
• Apr 6. Semantic science: what we can't do yet. Part 1, Part 2, Part 3, Part 4, Part 5, Part 6.
• Apr 8 - wrap up

Last updated: 2020-01-05, David Poole