Ruijie Wang

A replay-free GNN architecture effectively combats recency bias in class-incremental learning by routing nodes to experts via masked autoencoder reconstruction error, outperforming existing methods without task IDs.

Benign execution trajectories can leak proprietary agent skills, enabling reconstruction of high-value procedures without direct access to their underlying artifacts.

Achieve up to 52.5% compression in LLM chain-of-thought reasoning *while improving* accuracy by dynamically calibrating CoT length.