4 papers
From Deferral to Learning: Online In-Context Knowledge Distillation for LLM Cascades
Yu Wu, Shuo Wu, Ye Tao +2
Standard LLM cascades improve efficiency by deferring difficult queries from weak to strong models. However, these systems are typically static: when faced with repeated or semanti…
Learning to Help in Multi-Class Settings
Yu Wu, Yansong Li, Zeyu Dong +2
Deploying complex machine learning models on resource-constrained devices is challenging due to limited computational power, memory, and model retrainability. To address these limi…
Generalizing End-To-End Autonomous Driving In Real-World Environments Using Zero-Shot LLMs
Zeyu Dong, Yimin Zhu, Yansong Li +2
Traditional autonomous driving methods adopt a modular design, decomposing tasks into sub-tasks. In contrast, end-to-end autonomous driving directly outputs actions from raw sensor…
EAPCR: A Universal Feature Extractor for Scientific Data without Explicit Feature Relation Patterns
Zhuohang Yu, Ling An, Yansong Li +6
Conventional methods, including Decision Tree (DT)-based methods, have been effective in scientific tasks, such as non-image medical diagnostics, system anomaly detection, and inor…