Faisal Hamman
@FaisalHamman
Ph.D. student @UofMaryland | Algorithmic Fairness, Privacy & Explainability in Machine learning | Information theory.
ID:1662690685
http://faisalhamman.com 11-08-2013 14:16:47
725 Tweets
587 Followers
501 Following
📢Thrilled to announce that our paper titled 'Demystifying Local and Global Fairness Trade-offs in Federated Learning Using Partial Information Decomposition' has been accepted at #ICLR2024 This work was led by my PhD student Faisal Hamman #fairness #federatedlearning #reliableAI
📢 Paper: Robust Counterfactual Explanations for Neural Networks With Probabilistic Guarantees #ICML2023 #icml Faisal Hamman J.P. Morgan Please join at Exhibit Hall 1, Thu 27 Jul 10:30 a.m. HST.
If you are at #FAccT #Faact2023 , do attend Faisal Hamman’s talk on how querying for bias leaks protected attributes. Session 18: Privacy at 3:45 PM.
I won't be at #FAccT #FaccT2023 this year, but if you are, come say hello to Faisal Hamman! This paper is our first foray into privacy risks in algorithmic auditing - what information is leaked about people's identities when computing group-level disparities?
Happy to share our #FAccT2023 paper: Can Querying for Bias Leak Protected Attributes? arxiv.org/abs/2211.02139… (will be presented by Faisal Hamman, Joint work with Jiahao Chen)
This work was led by PhD student Faisal Hamman. Joint work with Saumitra Mishra and Daniele Magazzeni from J.P. Morgan
Read #FeaturePaper 'A Review of Partial Information Decomposition in Algorithmic Fairness and Explainability' from Sanghamitra Dutta and Faisal Hamman. mdpi.com/1099-4300/25/5…
#fairness
#explainability
#causality
#informationtheory
#uniqueinformation
In our paper, we show that querying for fairness metrics can leak the protected attributes of individuals. We propose a solution using differential privacy to prevent that.
Stop by our poster at the AFCP workshop #NeurIPS22 .
Room 392
4:20 PM CST
arxiv.org/abs/2211.02139