Course information

Course overview

What on your mind when presented the word “trustworthy”? In the era of machine learning, deep learning, and generative AI, the need for trustworthiness has never been so pressing. The key to ensure AI safety is to build a great wall of trustworthiness. In this course, we will cover several key topics related to trustworthy AI, including, but not limited to: robustness, generalization, responsibility, privacy, security, safety, and interpretability. By taking this course, students will form a basic understanding of essential concepts of trustworthiness, learn how to think beyond an existing AI system, develop key skills to deal with different untrustworthy conditions, read relevant paper, and even write their own research ideas. This course will have flexible forms, including lectures, guest lectures, paper reading, homework, and group projects.

Syllabus

Note 1: the following table contains all contents but may be subject to change since I could be traveling to other conferences or events, and/or the guest lecturers could be changed due to their private matters. Please be aware of my emails.

Note 2: each class lasts 80 minutes = 55min lecture + 5min buffer + 20min presentation. Each student is supposed to finish one presentation. For those who do not finish the presentations during the lecture sessions can present in the rest of the class (i.e., week 13, 14, or 15).

Week Topic/Content Resources/Suggested Reading Note
1 (08/27) L01: Introduction 1. KDD 2023 tutorial: Trustworthy Machine Learning
  1. Trustworthy AI: From Principles to Practices | | | 2 (09/01, 09/03) | L02: Basics of ML, DL, and Large Language Models | 1. On the Opportunities and Risks of Foundation Models
  2. What is Generative AI and How Does It Work? (video) | Homework 1 starts | | 3 (09/08, 09/10) | L03: Generalization 1—Transfer Learning | 1. Introduction to Transfer Learning
  3. A Comprehensive Survey on Transfer Learning | | | 4 (09/15, 09/17) | L04: Generalization 2—Domain Generalization | 1. Generalizing to Unseen Domains: A Survey on Domain Generalization
  4. IJCAI 2022 Tutorial on Domain Generalization
  5. DIVERSIFY: A General Framework for Time Series Out-of-Distribution Detection and Generalization
  6. AAAI 2024 tutorial on OOD generalization for time series | | | 5 (09/22, 09/24) | L05: Generalization 3—Weekly Supervised Learning | 1. SimCLR: A Simple Framework for Contrastive Learning of Visual Representations
  7. Masked autoencoders are scalable vision learners
  8. Momentum contrast for unsupervised visual representation learning
  9. FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
  10. FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling | Homework 1 due | | 6 (09/29, 10/01) | L06: Robustness and Security 1—Adversarial Robustness | 1. Explaining and Harnessing Adversarial Examples
  11. Towards Deep Learning Models Resistant to Adversarial Attacks
  12. BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain | Homework 2 starts | | 7 (10/06, 10/08) | L07: Robustness and Security 2—LLM Security | 1. Dyval: Dynamic evaluation of large language models for reasoning tasks
  13. The good, the bad, and why: Unveiling emotions in generative ai
  14. Understanding and Mitigating the Label Noise in Pre-training on Downstream Tasks | Fall break: 10/08-10/11 | | 8 (10/13, 10/15) | Guest Lecture: Prof. Kuan-Hao Huang from TAMU L07: Robustness and Security 2—LLM Security | 1. Backdoor learning: a survey | | | 9 (10/20, 10/22) | L08: Robustness and Security 3—Backdoor Attack L09: Privacy 1—AI Privacy | 1. Membership inference attacks against machine learning models
  15. Extracting Training Data from Large Language Models
  16. Regulating ChatGPT and Other Large Generative AI Models: A Privacy and Data Protection Perspective | Homework 2 due | | 10 (10/27, 10/29) | L10: Privacy 2—Federated Learning | 1. FedCLIP: Fast Generalization and Personalization for CLIP in Federated Learning
  17. MetaFed: Federated Learning among Federations with Cyclic Knowledge Distillation for Personalized Healthcare | Homework 3 starts | | 11 (11/03, 11/05) | L11: Privacy 3—Personalization | 1. Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization
  18. Personalization of Large Language Models: A Survey | 11/03: Election Day, No Class | | 12 (11/10, 11/12) | Guest lecture 2: Millicent Ochieng from Microsoft Research L11: Privacy 3—Personalization | 1. CulturePark: Boosting Cross-cultural Understanding in Large Language Models
  19. Promptrobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Prompts
  20. Constitutional AI: Harmlessness from AI Feedback | Homework 3 due Homework 4 starts | | 13 (11/17, 11/19) | L12: Human-AI Alignment

| 1. Fairness and Abstraction in Sociotechnical Systems | | | 14 (11/24, 11/26) | L13: Fairness | 1. Why Should I Trust You?": Explaining the Predictions of Any Classifier | 11/24-Remote; 11/25 - 11/29: Thanksgiving Break | | 15 (12/01, 12/03) | L14: Interpretability Final project preparation | | Homework 4 due |

Grading

Acknowledgements