MIT CSAIL6.882 Embodied Intelligence |
||
Spring 2020 |
||
This schedule is preliminary and subject to change as the term evolves.
| Date | Topics | Course Materials | Assignments | ||
| Week 1: Cognitive architectures and paradigms | |||||
| Tues 2/4/2019 | Domains, agent architecture, factory/wild | Russell & Norvig slides | |||
| Thurs 2/6/2019 | Classic papers |
Assigned readings: Shakey (video) (optional: Technical Report) Brooks, Flynn: Fast, Cheap, and Out of Control Sutton: Dyna |
Assignment #1 | ||
| Week 2: The problem of long-term decision making | |||||
| Tues 2/11/2019 | Deterministic domains: pddl planning, IW(k), power of heuristics, factoring |
Assigned readings: Bonet & Geffner: HSP Lipovetsky & Geffner: IW(k) Optional: Bandres, Bonet & Geffner: Planning with Pixels |
Assignment #2 | ||
| Thurs 2/13/2019 | Future-state uncertainty: value iteration, mcts, MPC / replanning |
Assigned readings: Kocsis & Szepesvari: Bandit-based Monte-Carlo Planning Yoon, Fern & Givan: FF-Replan: A Baseline for Probabilistic Planning Riedmiller: Neural Fitted Q Iteration |
Assignment #3 | ||
| Week 3: The problem of continuous action spaces | |||||
| Thurs 2/20/2019 | Motion planning, optimization |
Assigned readings: Kuffner & LaValle: RRT-Connect (animations) Mordatch et al.: Contact-Invariant Optimization (video and video) |
Assignment #4 | ||
| Week 4: The problem of situation awareness | |||||
| Tues 2/25/2019 | Perception as state estimation, object-oriented perception |
Assigned readings: Dense Object Nets Object-Oriented Prediction and Planning Optional: The feeling of success |
Assignment #5 | ||
| Thurs 2/27/2019 | Sensory representation learning |
Assigned readings: Large Scale Unsupervised Learning (cat face paper) Contrastive Predictive Coding Optional: Contrastive Multiview Coding |
Assignment #6 | ||
| Week 5: The problem of partial observability | |||||
| Tues 3/3/2019 | POMDPs: search and policies |
Assigned readings: POMDP Tutorial Belief-space Planning |
Assignment #7 | ||
| Thurs 3/5/2019 | POMDPs: Learning |
Assigned readings: RL in Robotics: A Survey (just review sections 2.2.2 and 2.3) Solving Deep Memory POMDPs Deep Recurrent Q-Learning for POMDPs |
Assignment #8 | ||
| Week 6: The problem of learning for long horizons | |||||
| Tues 3/10/2019 | Exploration, novelty search / curiosity, intrinsic motivation |
Assigned readings: Novelty Curiosity-Driven Exploration |
Assignment #9 | ||
| Thurs 3/12/2019 | HRL |
Assigned readings: Feudal RL h-DQN |
Assignment #10 | ||
| Week 7: classes canceled | |||||
| Thurs 3/26/2019 | Project | Assignment: Project handout (updated) |
|||
| Week 8: The problem of very little data | |||||
| Tues 3/31/2019 | Model-based intelligence | Assigned readings: Visual Foresight for Robotic Control World Models Optional: Building Machines That Learn and Think Like People (read Section 4, skim rest) |
Assignment #11 | ||
| Thurs 4/2/2019 | Meta-learning | Assigned readings: MAML Modular Meta-Learning Optional: RL^2 |
Assignment #12 | ||
| Week 9: The problem of nonstationarity | |||||
| Tues 4/7/2019 | Methodology | Assigned readings: Deep RL that Matters Assessing Generalization in Deep RL Optional: Machine Learning that Matters |
Final project proposal due Assignment #13 |
||
| Thurs 4/9/2019 | Cooperative Multiagent RL | Assigned readings: QMIX Learning to Communicate with Deep MARL Optional background: Comprehensive survey of MARL |
Assignment #14 | ||
| Week 10: The problem of adversaries | |||||
| Tues 4/14/2019 | Adversaries/Self-play | Assigned readings: Markov games TD-Gammon Optional: AlphaZero |
Assignment #15 | ||
| Thurs 4/16/2019 | Artificial life | Assigned readings: Karl Sims (evolving by competition) (video) (Optional: see also Karl Sims' prior paper on evolving agents without explicit competition) Hide and Seek (video) |
Assignment #16 | ||
| Week 11: Concrete problems I | |||||
| Tues 4/21/2019 | A general-purpose robot hand | Assigned Reading: Learning Dexterous In-Hand Manipulation (Movie) Optional: The follow-up paper on solving a Rubick's Cube |
|||
| Thurs 4/23/2019 | AI safety and social consequences | Assigned Reading: Self-Driving Cars and Engineering Ethics: The Need for a System Level Analysis (pdf will be posted on Piazza if you can't access this site) |
|||
| Week 12: Project milestone presentations | |||||
| Tues 4/28/2019 | Project milestone presentations | ||||
| Thurs 4/30/2019 | Project milestone presentations | ||||
| Week 13: Work on projects, no class meetings | |||||
| Week 14: Final project presentations | |||||
| Tues 5/12/2019 | Final project presentations | Final project due | |||