Shiqiang Wang

GneissWeb: Preparing High Quality Data for LLMs at Scale

Efficient Orthogonal Fine-Tuning with Principal Subspace Adaptation

Gradient Correction in Federated Learning with Adaptive Optimization

MESS+: Dynamically Learned Inference-Time LLM Routing in Model Zoos with Service Level Guarantees

Flick: Empowering Federated Learning with Commonsense Knowledge

RCCDA: Adaptive Model Updates in the Presence of Concept Drift under a Constrained Resource Budget

Vertical Federated Learning with Missing Features During Training and Inference

Dynamic Loss-Based Sample Reweighting for Improved Large Language Model Pretraining

Compressed Decentralized Learning with Error-Feedback under Data Heterogeneity