• Two papers accepted by EMNLP 2026!

    Two papers from our team were accepted by EMNLP 2026, including one main-conference paper and one Findings paper. Both works study computer-use agents in long-horizon real-world or clinical environments. Reference WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces Wanli Li, Bowen Zhou, Yunyao Yu, Zhou Xu, Yifan...

  • One paper accepted by COLM 2026!

    One paper from our team was accepted by COLM 2026. The work investigates why self-distillation can sometimes degrade the reasoning capability of large language models. Reference Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs? Jeonghye Kim, Xufang Luo, Minbeom Kim, Sangmook Lee, Dohyung Kim, Jiwon Jeon, Dongsheng Li,...

  • Four papers accepted by ICML 2026!

    Four papers from our team were accepted by ICML 2026. These works cover diffusion planning, neuro-inspired phase encoding, continual learning, and neural coding strategies. Reference Improving Diffusion Planners by Self-Supervised Action Gating with Energies Yuan Lu, Dongqi Han, Yansen Wang, Dongsheng Li. The Forty-third International Conference on Machine Learning (ICML)....

  • Two papers accepted by ACL 2026!

    Two papers from our team were accepted by ACL 2026, including one main-conference paper and one Findings paper. The works improve the efficiency of mixture-of-experts inference and full-parameter LLM fine-tuning. Reference Alloc-MoE: Budget-Aware Expert Activation Allocation for Efficient Mixture-of-Experts Inference Baihui Liu, Kaiyuan Tian, Wei Wang, Zhaoning Zhang, Linbo Qiao,...

  • Nine papers accepted by ICLR 2026!

    Nine papers from our team were accepted by ICLR 2026. These works span visual adaptation and interpretability, LLM agents, healthcare AI, neuromorphic learning, multimodal reasoning, reinforcement learning, and diffusion models. Reference Exploring Interpretability for Visual Prompt Tuning with Cross-layer Concepts Yubin Wang, Xinyang Jiang, De Cheng, Xiangqian Zhao, Zilong Wang,...

  • Six papers accepted by NeurIPS 2025!

    Six papers from our team were accepted by NeurIPS 2025, including one Spotlight paper. The works cover spiking Transformers, language-model learning and fine-tuning, visual instruction tuning, electrophysiology, and genomic foundation models. Reference Toward Relative Positional Encoding in Spiking Transformers Changze Lv, Yansen Wang, Dongqi Han, Yifei Shen, Xiaoqing Zheng, Xuanjing...

  • Two papers accepted by CVPR 2025!

    In this year’s CVPR conference, our team members have published two papers on vision-language model and domain generalization. Several team members will attend the conference in person and welcome everyone to stop by their posters for discussions. Reference Mitigating Hallucinations in Large Vision-Language Models via DPO: On-Policy Data Hold the...

  • Seven papers accepted by ICLR 2025!

    In this year’s ICLR conference, our team members have published seven papers in diffusion models, foundation models and LLM. Several team members will attend the conference in person and welcome everyone to stop by their posters for discussions. Reference How Do Large Language Models Understand Graph Patterns? A Benchmark for...