News

Sep 06, 2026 Our paper “OPSERVE: Opportunistic LLM Inference over Fragmented GPU Capacity in HPC Systems” was accepted to the AI on HPC Workshop (AIonHPC) at SC 2026.
Aug 14, 2026 New preprint: “Wrong but Useful: Trajectory Value Beyond Answer Correctness in Multi-Agent Messages”.
Aug 01, 2026 Our paper “BatchFlow: Shared Batch Reuse and Adaptive Scheduling for Multi-Job Training” was accepted to SC 2026 (19% acceptance rate).
May 13, 2026 New preprint: “PEML: Parameter-efficient Multi-Task Learning with Optimized Continuous Prompts”.
Apr 09, 2026 New preprint: “SOLAR: Communication-Efficient Model Adaptation via Subspace-Oriented Latent Adapter Reparametrization”.
Oct 03, 2025 New preprint: “Diffusion-Based, Data-Assimilation-Enabled Super-Resolution of Hub-height Winds”.
Mar 01, 2025 Two papers published in Performance Evaluation: “Enabling Scalable and Adaptive Machine Learning Training via Serverless Computing on Public Cloud” and “FedCust: Offloading Hyperparameter Customization for Federated Learning”.
May 07, 2024 Our paper “MalleTrain: Deep Neural Networks Training on Unfillable Supercomputer Nodes” was accepted to ICPE 2024.
Mar 01, 2023 Our paper “InfiniStore: Elastic Serverless Cloud Storage” was accepted to VLDB 2023.
Feb 24, 2020 Our paper “InfiniCache: Exploiting Ephemeral Serverless Functions to Build a Cost-Effective Memory Cache” was accepted to FAST 2020.