News
Papers, talks, milestones, and the occasional poem.
- PreprintFedeRage: Provably Convergent Agnostic Federated Learning under General Client Drift is now on arXiv.
- PreprintConstrained Learning with Universally Learnable Concept Classes is now on arXiv.
- PreprintCausal Longitudinal Prior-Fitted Networks for Counterfactual Outcome Prediction is now on arXiv.
- DegreeReceived my Master of Philosophy (M.Phil.) from Yale.
- TalkWeighted Stochastic Differential Equation to Implement Wasserstein–Fisher–Rao Gradient Flow was selected for a lightning talk at the Yale AI Symposium.
- TalkFedAVOT was accepted for presentation at the AI Symposium in Hartford.
- PaperFedAVOT was accepted at ICASSP 2026.
- PreprintWeighted Stochastic Differential Equation to Implement Wasserstein–Fisher–Rao Gradient Flow is now on arXiv.
- MilestoneDefended my prospectus — I am now a Ph.D. Candidate!
- PreprintUncertainty-Aware Generative Oversampling Using an Entropy-Guided Conditional Variational Autoencoder is now on arXiv.
- AwardConvergence of Agnostic Federated Averaging was accepted at IEEE CAMSAP 2025 and named a finalist in the Student Best Paper Contest.
- PreprintFedAVOT: Exact Distribution Alignment in Federated Learning via Masked Optimal Transport is now on arXiv.
- PreprintConvergence of Agnostic Federated Averaging is now on arXiv.
- DegreeReceived my Master of Science from Yale’s Department of Electrical & Computer Engineering.
- TeachingTaught a mini-course on Information Geometry, exploring geometric aspects of inference and divergence.
- AwardWon the Best Original Love Poem Award at Yale.
- FellowshipAwarded Yale’s Graduate Student Top-Up Fellowship ($1,250) for completing coursework with the highest overall GPA.
- ResearchBegan new research on risk-averse federated learning, combining CVaR and DRO frameworks for robust multi-client training.