Trust and Distrust in ML: Privacy, Verification and Robustness - Part 2

Overview

Delve into privacy, verification, and robustness challenges in machine learning with Shafi Goldwasser's exploration of trust issues in ML systems.

Syllabus

- Introduction to Trust and Distrust in Machine Learning -- Overview of trust issues in ML systems -- Key concepts of privacy, verification, and robustness - Privacy in Machine Learning -- Differential privacy -- Data anonymization techniques -- Privacy-preserving machine learning models -- Case studies on privacy failures - Verification in Machine Learning -- Formal verification methods for ML systems -- Testing and validating machine learning models -- Tools and techniques for model verification -- Real-world applications and scenarios - Robustness in Machine Learning -- Adversarial attacks and defenses -- Robustness testing and evaluation -- Designing robust ML systems -- Case studies on robustness challenges - Case Studies and Applications -- Examination of high-impact case studies -- Lessons learned from trust issues in past projects - Expert Guest Lecture: Shafi Goldwasser -- Deep dive into specific trust challenges -- Open Q&A session - Future Directions in Trust for ML -- Emerging trends and research areas -- Discuss unresolved challenges and potential solutions - Conclusion and Wrap-up -- Summary of key learnings -- Final thoughts on building trustworthy ML systems - Project: Building a Trustworthy ML Application -- Design and implement a small-scale ML system addressing privacy, verification, and robustness -- Present and critique projects in a peer review session

Taught by

Tags