3

Modeling with Homophily Driven Heterogeneous Data in Gossip Learning
Proposes how to handle label skew in decentralized learning where the users connect peer to peer.
Abhirup Ghosh, Cecilia Mascolo
DOI
Continual and Sliding Window Release for Private Empirical Risk Minimization
Releasing differentially private ML models for data streams.
Lauren Watson, Abhirup Ghosh, Benedek Rozemberczki, Rik Sarkar
Last updated on Jul 20, 2022
PDF
Evaluating the Feasibility of Gossip Learning on Android Phones (Workshop)
Evaluates if today’s smartphones are capable of running gossip learning.
Issam Nedjaï, Young D. Kwon, Abhirup Ghosh, Cecilia Mascolo
Last updated on Jul 20, 2022
In-Network Approximate and Efficient Spatiotemporal Range Queries on Moving Objects
Distributed query on moving objects using tools from differential topology and Machine learning.
Guang Yang, Abhirup Ghosh, Thomas Heinis
Last updated on Jul 20, 2022
Investigating Domain agnostic Performance in Activity Recognition using Accelerometer Data (Workshop)
Generalizability of HAR models to unseen data distribution.
Api Hasthanasombat, Abhirup Ghosh, Cecilia Mascolo
Last updated on Jul 20, 2022
Less Data Can Be Better for Domain Generalization
Generalization to unknown test distribution improves when a model is
Api Hasthanasombat, Abhirup Ghosh, Cecilia Mascolo
Last updated on Jul 20, 2022