Instructor : Momotaz Begum, Assistant Professor, Computer Science, UNH; mbegum@cs.unh.edu
Classes : MW: 2:10 PM– 3:30 PM
Location : Kingsbury N 233
Office hours : T: 10:00 AM - 12:00 PM: zoom
Week 1:
Class introduction
Week 2: Survey on LfD
Recent Advances in Robot Learning from Demonstrations (Noushad, Alex)
Week 3: Survey on visual imitation learning
Recent Advances in Imitation Learning from Observations (Shatanu)
Sample-efficient Adversarial Imitation Learning from Observation (Stephen)
Imitation Learning: A Survey of Learning Methods
Imitation Learning: Progress, Taxonomies and Challenges
Week 4 -5: Continuous/discreate control from raw observation (BC+RL+IRL)
One-Shot Visual Imitation Learning via Meta-Learning (Palash)
Time-Contrastive Networks: Self-Supervised Learning from Video (Stephen)
Third-Person Visual Imitation Learning via Decoupled Hierarchical Controller
Imitation from observation: Learning to imitate behaviors from raw video via context translation
Zero-shot visual imitation learning
Robust Behavior Cloning with Adversarial Demonstration Detection
A paper on inverse-dynamic model:
Behavioral cloning from observation
Week 6: Understanding task structure (or a plan) learning from video
Graph-Structured Visual Imitation
Automated Generation of Robotic Planning Domains from Observations
Neural Task Graphs: Generalizing to Unseen Tasks from a Single Video Demonstration
Efficient and Interpretable Robot Manipulation with Graph Neural Networks
Week 7: Covariate Shift in VIL
A reduction of imitation learning and structured prediction to no-regret online learning
Back to the Manifold: Recovering from Out-of-Distribution States
The Surprising Effectiveness of Representation Learning for Visual Imitation
Week 8: Language and vision for IL
BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learning
Do As I Can, Not As I Say: Grounding Language in Robotic Affordances
(Optional) Week 9: Correspondence problem in VIL
AVID: Learning Multi-Stage Tasks via Pixel-Level Translation of Human Videos
Learning by Watching: Physical Imitation of Manipulation Skills from Human Videos
Week 9 - Week 14: Project work begins and we study related papers
Week 15: Project presentation