min null 1
Multi-Agent Learning with Heterogeneous Linear Contextual Bandits
UCB, wherein agents cooperatively minimize the group regret under the coordination of a central server.
On Optimal Caching and Model Multiplexing for Large Model Inference
By combining a caching algorithm, namely Greedy Dual Size with Frequency (GDSF) or Least Expected Cost (LEC), with a model multiplexer, we achieve optimal rates in both offline and online settings.
Precise Asymptotic Generalization for Multiclass Classification with Overparameterized Linear Models
We fully resolve the conjecture posed in Subramanian et al. (2022),
Public-data Assisted Private Stochastic Optimization: Power and Limitations
We study the limits and capability of public-data assisted differentially private (P A-DP) algorithms. Specifically, we focus on the problem of stochastic convex optimization (SCO) with either labeled or unlabeled public data.
2Direction: Theoretically Faster Distributed Training with Bidirectional Communication Compression
We consider distributed convex optimization problems in the regime when the communication between the server and the workers is expensive in both uplink and downlink directions.