📌 2026 So Far — 8 Papers Accepted

A quick recap of what our lab has accepted so far in 2026:

1. Online Conformal Prediction with Adversarial Semi-bandit Feedback via Regret Minimization — ICLR'26
2. Ruby: Unmasking Unsafe Rust in Stripped Binaries via Machine Learning — DSN'26
3. LLM Watermark Evasion via Bias Inversion — ICML'26
4. Towards Functional Correctness of Large Code Models with Selective Generation — ICML'26
5. The Illusion of Rust Safety: Detecting Modular Unsafe Functions with LLMs — CCS'26
6. SAUL: Sharpness-Aware Augmented-Lagrangian Unlearning — EMNLP Findings'26 (🏆 Best Paper Award from CKAIA)
7. Can We Stop Malicious AI? KILLBENCH: A Benchmark for External AI Kill Switch Feasibility — EMNLP Findings'26
8. The Interplay of Harness Design and Post-Training in LLM Agents — EMNLP Findings'26

Also serving as an area chair for NeurIPS'26 and ICLR'27. Congrats to everyone involved — more to come!