24-784 Trustworthy AI Autonomy (TAIAT), Spring 2022

Instructor — Prof. Ding Zhao
Teaching assistants — Wenhao Ding, Zuxin Liu, Jiayin Xia, Zijian Guo, Sameer Bharadwaj, Yihan He
Course linksCanvas site, Course webpage

Course objective — Machine learning is a fast developing field. This is a big opportunity for young generation but also a big challenge as one may get lost in the pyramid of technical news and papers. This course aims to achieve two goals:
  • Help students have a high-level understanding of the fast-developing TAIAT area, quickly grasp key concepts and skills, and get familiar with the latest technological tools in the early phase of their graduate program, so that they could focus on a direction to establish their expertise;
  • Develop full-cycle research capabilities including paper reviews, writing research proposals, milestone reports, academic papers, conference-style paper reviews, making presentations, and serving as an academic critic.

Lecture schedule for Spring 2022

Week Day, Date Lec# Topic
1 Tuesday, Jan 18, 2022 1 Overview, autonomy framework, trustworthy autonomy [Slide]
1 Thusrday, Jan 20, 2022 2 Deep learning basics, vision models [Slide]
2 Tuesday, Jan 25, 2022 3 Latent space visualization, explanability [Slide]
2 Thursday, Jan 27, 2022 4 Security attacks: poisoning, evasion, FGSM, robust physical attack [Slide]
3 Tuesday, Feb 1, 2022 5 Robustness-Adversarial and defensive ML: randomization, robust AI, certification [Slide]
3 Thursday, Feb 3, 2022 6 Model-free decision making: imitation learning, reinforcement learning, Q learning [Slide]
4 Tuesday, Feb 8, 2022 7 Model-free Deep RL: REINFORCE, Actor-Critic [Slide]
4 Thursday, Feb 10, 2022 8 Model-based Deep RL: MPC [Slide]
5 Tuesday, Feb 15, 2022 9 Adversarial AI
5 Thursday, Feb 17, 2022 10 Gaussian processes: GP[Slide]
6 Tuesday, Feb 22, 2022 11 Safety: CMDP, Lagrangian-based Method (TRPO-lag, PPO-lag), Constrained Optimizationl [Slide]
6 Thursday, Feb 24, 2022 12 RL for real world autonomy (and model based RL)
7 Tuesday, Mar 1, 2022 13 Safety: CMDP, Lagrangian-based Method (TRPO-lag, PPO-lag), Constrained Optimization [Slide]
7 Thursday, Mar 3, 2022 14 Safety: reachability, Control Lyapunov, barrier function [Slide]
8 Tuesday, Mar 8, 2022 No class
8 Thursday, Mar 10, 2022 No class
9 Tuesday, Mar 15, 2022 15 Certification: overview, digital twin simulation, safety critical scenario generation [Slide]
9 Thursday, Mar 17, 2022 16 Safe RL
10 Tuesday, Mar 22, 2022 17 Digital twin - data-driven: VAE, GAN, and Flow [Slide]
10 Thursday, Mar 24, 2022 18 Digital twin - adversarial: worst-case, IS, splitting [Slide]
11 Tuesday, Mar 29, 2022 19 Scenario generation, evaluation and certification
11 Thursday, Mar 31, 2022 20 Generalization: Working with real world robots, domain randominzation, DDPG, SAC [Slide]
12 Tuesday, Apr 5, 2022 21 Generalization: Nonstationary environment: delay, RARL, meta learning, NP [Slide]
12 Thursday, Apr 7, 2022 No class
13 Tuesday, Apr 12, 2022 22 Milestone review
13 Thursday, Apr 14, 2022 23 Generalization: hierachical AI, life long learning, DPGP [Slide]
14 Tuesday, Apr 19, 2022 24 Generalization
14 Thursday, Apr 21, 2022 25 Human centricity: Privacy, fairness [Slide]
15 Tuesday, Apr 26, 2022 26 Invited speaker: Rahul Mangharam
15 Thursday, Apr 28, 2022 27 Alumni session: presentation