MS student, Jilin University
1 paper at NeurIPS 2025
We propose an end-to-end autonomous driving framework based on a unified diffusion-regression-classification policy, achieving state-of-the-art performance on both the CARLA and NAVSIM benchmarks.
