Avishek Ghosh

Label-flipping attacks on federated GANs can skew generation distributions significantly while remaining undetectable by traditional label-agnostic metrics.

Optimizing arm selection in multi-agent bandit settings can eliminate dependence on the number of arms, achieving optimal regret even in adversarial environments.

Gradient EM provably converges to optimal solutions for fitting mixtures of *any* parametric function with strongly convex losses, extending beyond just linear regression in the agnostic setting.