activity
20242026
collaborators

5 papers

cs.AI2026

When Shared Rollouts Fail in Defensive Driving Evaluation: A NAVSIM Score Basis Audit

Ziang Wei, Minjun Yu, Zheyuan Lai +2

Defensive driving scores are useful only when they preserve distinctions between policies that observe surrounding actors and those that do not. Re-simulation benchmarks may use re…

math.PR2026

Information-theoretic coordinate subset and partition selection of multivariate Markov chains via submodular optimization

Zheyuan Lai, Michael C. H. Choi

We study the problem of optimally projecting the transition matrix of a finite ergodic multivariate Markov chain onto a lower-dimensional state space, as well as the problem of fin…

math.PR2026

Information-theoretic minimax and submodular optimization algorithms for multivariate Markov chains

Zheyuan Lai, Michael C. H. Choi

We study an information-theoretic minimax problem for finite multivariate Markov chains on -dimensional product state spaces. Given a family of $…

cs.LG2025

PriM: Principle-Inspired Material Discovery through Multi-Agent Collaboration

Zheyuan Lai, Yingming Pu

Complex chemical space and limited knowledge scope with biases holds immense challenge for human scientists, yet in automated materials discovery. Existing intelligent methods reli…

cs.CV2024

ESP: Extro-Spective Prediction for Long-term Behavior Reasoning in Emergency Scenarios

Dingrui Wang, Zheyuan Lai, Yuda Li +5

Emergent-scene safety is the key milestone for fully autonomous driving, and reliable on-time prediction is essential to maintain safety in emergency scenarios. However, these emer…