Published in arXiv preprint, 2026
We propose CAST, an answer-free self-distillation method for GRPO-style RLVR that shapes token-level advantages according to trajectory correctness.
Recommended citation: Yang Li, Gongle Xue, Yijia Guo, Yuheng Yuan, Liwen Hu, Lei Ma. "CAST: Non-Privileged Clipped Asymmetric Self-Teaching with Advantage Flipping for GRPO." arXiv preprint arXiv:2606.00172, 2026.
Download Paper
Published in AAAI 2026, 2026
We propose splats in splats, the first 3DGS steganography framework that embeds 3D content in 3DGS itself without modifying any attributes, via importance-graded SH coefficient encryption and autoencoder-assisted opacity mapping.
Recommended citation: Yijia Guo, Wenkai Huang, Yang Li, Gaolei Li, Hang Zhang, Liwen Hu, Jianhua Li, Tiejun Huang, Lei Ma. "Splats in Splats: Robust and Effective 3D Steganography towards Gaussian Splatting." Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2026.
Download Paper
Published in ECCV 2026, 2025
We propose Diff-SPK, a diffusion-based reconstruction method that leverages generative priors to recover high-quality images from low-light spike streams.
Recommended citation: Liwen Hu, Yang Li, Mianzhi Liu, Yijia Guo, Shenghao Xie, Ziluo Ding, Tiejun Huang, Lei Ma. "Seeing the Unseen in Low-light Spike Streams." European Conference on Computer Vision (ECCV), 2026.
Download Paper