Research

Publications

Research on efficient machine learning systems, long-context inference, and multi-agent systems.

2026

EMAS: Stabilizing Multi-Agent System Evolution through Evidence-Guided Revision

Chao Fei, Qingyi Si, Kaihua Liang, Yanghua Xiao, Panos Kalnis, Hongcheng Guo

arXiv preprint · 2026

Evolves multi-agent prompts and topology from recurring evidence-guided diagnoses, accepting revisions only after paired validation improves accuracy or cost.

Exploring a Layer-Wise Design Space for KV Cache Eviction

Chao Fei, Kaihua Liang, Hanzhi Hu, Hongcheng Guo, Jian Weng, Marco Canini, Panos Kalnis

arXiv preprint · 2026

Explores layer-wise composition of existing KV-cache eviction methods, showing that heterogeneous routing and profile-guided placement improve long-context quality under matched cache budgets.

CHESS: Context-aware Hierarchical Efficient Semantic Selection for Long-Context LLM Inference

Chao Fei, Guozhong Li, Chenxi Liu, Panos Kalnis

arXiv preprint · 2026

Combines context-aware hierarchical KV selection with coarse-grained system support to retain quality while accelerating long-context decoding.