Jiaxing Wang

profile picture

Jiaxing Wang (王家兴)

Researcher, JD.com

Beijing, China

Email: wjiaxing94 [at] gmail.com

I am currently a researcher at JD.com. I obtained my Ph.D. degree from National Laboratory of Pattern Recognition from CASIA, supervised by Prof. Jian Cheng, and the B.Eng. Degree from North China Electric Power University.

My research focuses on Data-centric AI and AI Infrastructure, with a specific emphasis on data mixture optimization, data selection, and data synthesis to construct a self-evolving data flywheel, as well as efficient model training and inference. These techniques have been successfully deployed in production within JD Retail’s commodity understanding, yielding substantial business impact.

news

Oct 13, 2026 Our papers “BLADE: Scalable Bi-level Adaptive Data Selection for LLM Training” and “The Hidden Power of Scaling Factor in LoRA Optimization” are accepted by NeurIPS 2026.
Aug 21, 2026 Our paper “FlowTrain: Flow-Based Decoupled Training for Industrial-Grade Vision-Language Models” and “MixEmb: Importance-Aware Quantization and Caching Co-Design for Industrial Decentral- ized Embedding Training” are accepted by EMNLP 2026 Industrial Track.
Jun 05, 2026 We present JD Oxygen AI Item Center (Oxygen AIIC) V1, an industrial-scale LLM/VLM-centric solution for item understanding, management, and applications.
Sep 05, 2025 Our paper “The Primacy of Magnitude in Low-Rank Adaptation” is accepted by NeurIPS 2025 (Spot Light).
Sep 05, 2025 Our paper “TANDEM: Bi-Level Data Mixture Optimization with Twin Networks” is accepted by NeurIPS 2025.

selected publications

  1. NeurIPS
    BLADE: Scalable Bi-level Adaptive Data Selection for LLM Training
    Jiaxing Wang*, Deping Xiang*, Jin Xu, and 9 more authors
    In The Fortieth Annual Conference on Neural Information Processing Systems, 2026
    PDF
  2. ATC
    Rollout-Training Co-Design for Efficient LLM-Based Multi-Agent Reinforcement Learning
    Zhida Jiang, Zhaolong Xing, Jiawei Lu, and 13 more authors
    In 2026 USENIX Annual Technical Conference, 2026
    PDF
  3. NeurIPS
    TANDEM: Bi-Level Data Mixture Optimization with Twin Networks
    Jiaxing Wang*, Deping Xiang*, Jin Xu, and 8 more authors
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025
    PDF
  4. NeurIPS
    The Primacy of Magnitude in Low-Rank Adaptation
    Zicheng Zhang, Haoran Li, Yifeng Zhang, and 5 more authors
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025
    PDF
  5. ACL
    Beyond Logits: Aligning Feature Dynamics for Effective Knowledge Distillation
    Guoqiang Gong*, Jiaxing Wang*, Jin Xu, and 8 more authors
    In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2025
    PDF
  6. AAAI
    Patch-aware sample selection for efficient masked image modeling
    Zhengyang Zhuge, Jiaxing Wang, Yong Li, and 3 more authors
    In Proceedings of the AAAI Conference on Artificial Intelligence, 2024
    PDF
  7. ICLR
    DynaMS: Dyanmic Margin Selection for Efficient Deep Learning.
    Jiaxing Wang, Yong Li, Jingwei Zhuo, and 7 more authors
    In ICLR, 2023
    PDF
  8. AAAI
    Dpnas: Neural architecture search for deep learning with differential privacy
    Anda Cheng, Jiaxing Wang, Xi Sheryl Zhang, and 3 more authors
    In Proceedings of the AAAI conference on artificial intelligence, 2022
    PDF
  9. NeurIPS
    Revisiting parameter sharing for automatic neural channel number search
    Jiaxing Wang*, Haoli Bai*, Jiaxiang Wu, and 5 more authors
    Advances in Neural Information Processing Systems, 2020
    PDF Code
  10. AAAI
    M-nas: Meta neural architecture search
    Jiaxing Wang, Jiaxiang Wu, Haoli Bai, and 1 more author
    In Proceedings of the AAAI conference on artificial intelligence, 2020
    PDF
  11. JSTSP
    Bayesian automatic model compression
    Jiaxing Wang, Haoli Bai, Jiaxiang Wu, and 1 more author
    IEEE Journal of Selected Topics in Signal Processing, 2020
    PDF
  12. ICML
    RaFM: Rank-aware factorization machines
    Xiaoshuang Chen, Yin Zheng, Jiaxing Wang, and 2 more authors
    In International Conference on Machine Learning, 2019
    PDF
  13. TNNLS
    ECBC: Efficient convolution via blocked columnizing
    Tianli Zhao, Qinghao Hu, Xiangyu He, and 4 more authors
    IEEE Transactions on Neural Networks and Learning Systems, 2021
    PDF

Services

  • Program Committee Member / Reviewer: NeurIPS 21-26, ICML 22-26, ICLR 22-26, ACL 26
  • Top Reviewer: NeurIPS 2026