Wu Zhaoxuan

DAVINZ-DataValuation. Training-free data valuation on deep neural network applications. (ICML-2022)

Incentive-Aware-FL. A novel algorithm that distributes training-time model rewards to incentivize client contributions for federated learning. (ICLR-2024)

EASE-Prompt-Optimization. An efficient automated exemplar selection method that uses a neural bandit algorithm to optimize the set of exemplars for in-context learning while accounting for exemplar ordering.

Tetris. Tetris optimizes the total throughput of batch speculative decoding in multi-request settings by strategically selecting draft tokens for verification.

VolumeBased-DataValuation. Validation free and replication robust volume-based data valuation method. (NeurIPS-2021)

Trusted-Maximizers-Entropy-Search-BO. Official implementation of the UAI 2021 paper "Trusted-Maximizers Entropy Search for Efficient Bayesian Optimization".