Date | Topic |
---|---|
Mar 12 -- Apr 7 | Papers presentations Piazza post |
Tue Mar 10 | No class (Ottawa trip) |
Feb 11 - Mar 5 |
Policy Gradient methods PolicyGradient.pdf ModelBasedRL ModelBasedRL.pdf Imitation Learning ImitationLearning.pfd Sim-to-real sim2real.pdf Hindsight Experience Replay (HER) and Advantage Actor Critic (AAC) HER-and-AAC.pdf |
Thu Feb 6 |
RL slides up until now: RL-slides-Feb6-2020.pdf Deep Q-learning (DQN); Deep Deterministic Policy Gradients (DDPG) Assignment 3 (Tabular Q learning) is out. https://github.com/UBCMOCCA/CPSC533V_a3 (view and edit using: python notebook hw3.ipynb) A good Jupyter notebook tutorial Due next Friday, Fri Feb 14, by email to Michiel (van@cs) |
Tue Feb 4 | HJB equations HJB methods thesis level sets toolbox (Ian Mitchell) |
Thu Jan 30 | Basic RL: RL-2.pdf returns, discount factor, policy cloning, episodic tasks, simple deterministic graph MDP, generalized policy iteration, state value function, bootstrapping, prioritized sweeping, experience tuples, stochastic dynamics function, Bellman backups, Bellman backup diagram, model-based vs model-free, policy objective functions A2 out (due Fri Feb 7) car drifting control: video paper 2D manipulation using RL: project web page |
Tue Jan 28 | More dynamics: constrained rigid body dynamics: general form in maximal coords, constraint stabilization, reduced coordinates, modeling collision impulses physics-two-body-complete.pdf physics-collisions.pdf |
Thu Jan 23 | Two rigid bodies with a constraint |
Tue Jan 21 | Rigid body dynamics (continued) physics-rigid-body-complete.pdf |
Thu Jan 16 | Rigid Body Dynamics physics-rigid-body-blank.pdf |
Tue Jan 14 | Introduction to Control control-blank.pdf a1.pdf |
Thu Jan 9 | Example of movement skills (video links to be posted); RL-intro.pdf |
Tue Jan 7 | Course overview; introductions; background survey. motion-notations.pdf |
(various) |
Representations of movement motion-notations.pdf prehistoric animated paintings (youtube) prehistoric animated paintings: paper (PDF) animated point light horse display visualizing human motion; motion sculptures Pixar Zoetrope: Animal movement leaping goat chicken head tracking: falcon cat stepping snake climbing tree goats on dam Scaling of sensorimotor control in terrestrial mammals Human movement baseball catching strategies optimizing energetic cost during walking unpowered exoskeleton Passive dynamics passive dynamic walking passive juggling bicycle stability split belt energy harvesting wheel-with-mass diesel governor steam governor Robotics Komura lab: high-speed robotics fast running robot triple pendulum baking robot: 2'31 in parkour atlas RL for movement NeurIPS 2019 learn to walk |