Part II project proposals (2019-20)
1) Anonymized Data Linkability via Approximate Graph Matching
2) Active Learning on Graph Edge Space
3) Countering Adversarial Attacks in Distributed Learning via User-Level Privacy