@article{guo2026atriadawn,title={{Atria Dawn: The Dawn of Agentic Superintelligence}},author={Guo, Honglin and Gui, Tao and Cai, Kun and Chen, Haodong and Chen, Yicheng and Dong, Guanting and Ge, Qiming and Hu, Yuyang and Huang, Zixian and Jin, Jiajie and Lam, Alexander and Li, Yining and Lin, Jiahang and Liu, Yanjiang and Lu, Xinyu and Lv, Haijun and Ma, Zerun and Shang, Junlin and Su, Qisheng and Wang, Guoqiang and Wang, Rui and Wang, Zhecan and Xiang, Hao and Xie, Xinchen and Xing, Shuhao and Xing, Xiaoyu and Xu, Wanghan and Yang, Xinyu and Yang, Yajie and Zhao, Chengfeng and Zhao, Haoran and Zhao, Penghao and Zhou, Ruojun and Zhou, Yunhua and Zhu, Dongsheng and Zou, Yicheng and Cai, Qiye and Che, Xinmeng and Chen, Jiabei and Chen, Jiahao and Chen, Jiayi and Chen, Yujia and Cui, Lizhi and Dai, Youheng and Deng, Xin and Dong, Yi and Dou, Shihan and Gu, Chenya and Guo, Xu and Han, Ding and Hao, Feiyang and He, Haotan and Hou, Jie and Hu, Binze and Hu, Zijian and Huang, Junhao and Jiang, Huicheng and Jiang, Jiazhen and Jiang, Shufan and Kuang, Jiahao and Lai, Bowen and Li, Bo and Li, Jiaqiang and Li, Peng and Li, Qilong and Li, Zhuoqun and Liu, Jiaxiang and Liu, Shuainan and Liu, Tong and Liu, Yi and Lu, Zhonghang and Luo, Jianwen and Luo, Yanyi and Lv, Huijie and Ma, Ningsheng and Min, Houcheng and Pan, Chengjun and Peng, Qiyuan and Peng, Xiaoxuan and Qian, Jianmin and Qiu, Jiantao and Ren, Wanying and Sha, Huayu and Shan, Jifei and Shang, Zixin and Shao, Bing and Sheng, Zhuohui and Shi, Jiayang and Shu, Yang and Simayi, Aierpanjiang and Song, Sirui and Song, Yuxiao and Sun, Zhe and Sun, Zhichao and Tan, Wenzhe and Tian, Wenhui and Tian, Zhongbo and Wang, Hanchen and Wang, Pengbo and Wang, Rui and Wang, Yiding and Wang, Yuhui and Xi, Zhiheng and Xu, Caijun and Xu, Chao and Xu, Yongfeng and Yang, Xiaolei and Yang, Zhixiong and Yao, Qian and Yi, Shihong and Ying, Yuankai and Yu, Jia and Yuan, Dingbo and Yuan, Hao and Yuan, Junjie and Zhang, Bo and Zhang, Caixian and Zhang, Qiuyinzhe and Zhao, Jiyuan and Zhao, Ying and Zheng, Pujun and Zhong, Xiaoxue and Zhou, Xiaohao and Zhou, Xinyu and Zhu, Guanru and Zhu, Yulun and Lu, Yaojie and Ji, Tao and Lin, Hongyu and Zhu, Yutao and Cao, Pengfei and He, Guoxiu and Han, Xianpei and He, Ben and Dou, Zhicheng and Liu, Kang and Zhang, Qi and Sun, Le and Zhao, Jun and Wen, Ji-Rong and Huang, Xuanjing and Jiang, Yu-Gang and Zhou, Bowen},journal={arXiv preprint arXiv:2609.15818},year={2026},month=sep,url={https://arxiv.org/abs/2609.15818},}
Preprint
ICAE-Bench: Evaluating Coding Agents as Interactive Project Builders
Zhongyuan Peng, Dan Huang, Chuyu Zhang, Caijun Xu, Changyi Xiao, Shibo Hong, David Lo, Lin Qiu, Xuezhi Cao, Jiyuan He, and 1 more author
@article{peng2026icae,title={{ICAE-Bench: Evaluating Coding Agents as Interactive Project Builders}},author={Peng, Zhongyuan and Huang, Dan and Zhang, Chuyu and Xu, Caijun and Xiao, Changyi and Hong, Shibo and Lo, David and Qiu, Lin and Cao, Xuezhi and He, Jiyuan and Cao, Yixin},journal={arXiv preprint arXiv:2607.21217},year={2026},month=jul,url={https://arxiv.org/abs/2607.21217},}
Preprint
DenoiseRL: Bootstrapping Reasoning Models to Recover from Noisy Prefixes
Caijun Xu, Changyi Xiao, Zhongyuan Peng, and Yixin Cao
DenoiseRL is a reinforcement learning framework that replaces external supervision with recovery-oriented optimization over failures from weak models. It learns directly from incorrect reasoning traces by conditioning the policy on noisy prefixes and optimizing the on-policy continuation, improving reasoning performance, exploration efficiency, and self-corrective behavior without stronger teacher models or heavily curated difficult datasets.
@article{xu2026denoiserl,title={{DenoiseRL: Bootstrapping Reasoning Models to Recover from Noisy Prefixes}},author={Xu, Caijun and Xiao, Changyi and Peng, Zhongyuan and Cao, Yixin},journal={arXiv preprint arXiv:2605.28421},year={2026},month=may,url={https://arxiv.org/abs/2605.28421},}
Preprint
Reinforcement Learning with Conditional Expectation Reward
@article{xiao2026conditional,title={{Reinforcement Learning with Conditional Expectation Reward}},author={Xiao, Changyi and Xu, Caijun and Cao, Yixin},journal={arXiv preprint arXiv:2603.10624},year={2026},month=mar,url={https://arxiv.org/abs/2603.10624},}
Preprint
CoDiQ: Test-Time Scaling for Controllable Difficult Question Generation
Zhongyuan Peng, Caijun Xu, Changyi Xiao, Shibo Hong, Eli Zhang, Stephen Huang, and Yixin Cao
@article{peng2026codiq,title={{CoDiQ: Test-Time Scaling for Controllable Difficult Question Generation}},author={Peng, Zhongyuan and Xu, Caijun and Xiao, Changyi and Hong, Shibo and Zhang, Eli and Huang, Stephen and Cao, Yixin},journal={arXiv preprint arXiv:2602.01660},year={2026},month=feb,url={https://arxiv.org/abs/2602.01660},}
ACL’26 Findings
SCALER: Synthetic Scalable Adaptive Learning Environment for Reasoning
Caijun Xu, Changyi Xiao, Zhongyuan Peng, Xinrun Wang, and Yixin Cao
In Findings of the Association for Computational Linguistics: ACL 2026, Jan 2026
@inproceedings{xu2026scaler,title={{SCALER: Synthetic Scalable Adaptive Learning Environment for Reasoning}},author={Xu, Caijun and Xiao, Changyi and Peng, Zhongyuan and Wang, Xinrun and Cao, Yixin},booktitle={Findings of the Association for Computational Linguistics: {ACL} 2026},year={2026},month=jan,url={https://arxiv.org/abs/2601.04809},}
2024
CIKM’24
Exploring High-Order User Preference with Knowledge Graph for Recommendation
Caijun Xu, Fuwei Zhang, Zhao Zhang, Fuzhen Zhuang, and Rui Liu
In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, Jan 2024
@inproceedings{xu2024exploring,title={{Exploring High-Order User Preference with Knowledge Graph for Recommendation}},author={Xu, Caijun and Zhang, Fuwei and Zhang, Zhao and Zhuang, Fuzhen and Liu, Rui},booktitle={Proceedings of the 33rd {ACM} International Conference on Information and Knowledge Management},pages={4138--4142},year={2024},url={https://doi.org/10.1145/3627673.3679921},}