Yaqi Xie

Postdoctoral Fellow of Carnegie Mellon University (CMU).

I am currently a postdoctoral researcher at Carnegie Mellon University (CMU) in the Robotics Institute advised by Prof. Katia Sycara. Before that, I received my Ph.D. in Computer Science from National University Singapore (NUS), where I was fortunated to be advised by Prof. Harold Soh.

My work centers on Human-AI Synergy through neural-symbolic fusion, bridging the complementary strengths of symbolic reasoning and deep learning to create adaptive, interpretable intelligence systems. I aim to integrate structured domain knowledge with neural architectures to enhance robustness, efficiency, and trustworthiness in dynamic environments. My research applications span perception, decision-making, and generative models.

In the long run, I strive to build adaptive AI agents that not only learn from data but also incorporate human knowledge and feedback, work seamlessly with humans, and align with human needs to enhance everyday life.

Feel free to contact me if you are interested in my research and want to discuss relevant research topic or potential collaborations!

News

Oct 2025 I’m excited to be on the faculty job market this year and would love to connect!
Feb 2023 I joined Carnegie Mellow University (CMU) as a postdoctoral fellow.

Selected Publications

For a complete list of publications, please visit my Google Scholar page.

  1. Unifying Deep Predicate Invention with Pre-trained Foundation Models
    Qianwei Wang, Bowen Li, Zhanpeng Luo, Yifan Xu, Alexander Gray, Tom Silver, Sebastian Scherer, Katia Sycara, and Yaqi Xie
    In Preprint, 2026
  2. Evolving Contextual Safety in Multi-Modal Large Language Models via Inference-Time Self- Reflective Memory
    Ce Zhang, Jinxi He, Junyi He, Katia Sycara, and Yaqi Xie
    In Conference on Computer Vision and Pattern Recognition (CVPR), 2026
    CVPR
  3. pySpatial: Generating 3D Visual Programs for Zero-Shot Spatial Reasoning
    Zhanpeng Luo, Ce Zhang, Silong Yong, Cunxi Dai, Qianwei Wang, Haoxi Ran, Guanya Shi, Katia Sycara, and  Yaqi Xie
    In International Conference on Learning Representations (ICLR), 2026
    ICLR
  4. Thought Communication in Multiagent Collaboration
    Yujia Zheng, Zhuokai Zhao, Zijian Li, Yaqi Xie, Mingze Gao, Lizhu Zhang, and Kun Zhang
    In Advances in Neural Information Processing Systems (NeurIPS), 2025
    NeurIPSSpotlight (3%)
  5. Adaptively Coordinating with Novel Partners via Learned Latent Strategies
    Benjamin Li, Shuyang Shi, Lucia Romero, Huao Li, Yaqi Xie, Woojun Kim, Stefanos Nikolaidis, Charles Michael Lewis, Katia Sycara, and Simon Stepputtis
    In Advances in Neural Information Processing Systems (NeurIPS), 2025
    NeurIPSBest Paper at RSS GenAI-HRI
  6. Self-Correcting Decoding with Generative Feedback for Mitigating Hallucinations in Large Vision-Language Models
    Ce Zhang, Zifu Wan, Zhehan Kan, Martin Q. Ma, Simon Stepputtis, Deva Ramanan, Russ Salakhutdinov, Louis-Philippe Morency, Katia Sycara, and Yaqi Xie
    In International Conference on Learning Representations (ICLR), 2025
    ICLR
  7. Let Me Help You! Neuro-Symbolic Short-Context Action Anticipation
    Sarthak Bhagat, Samuel Li, Joseph Campbell, Yaqi Xie, Katia Sycara, and Simon Stepputtis
    IEEE Robotics and Automation Letters (RA-L), 2024
    RA-L, ICRABest Paper at ICRA Nonverbal Cues
  8. ShapeGrasp: Zero-Shot Task-Oriented Grasping with Large Language Models through Geometric Decomposition
    Samuel Li, Sarthak Bhagat, Joseph Campbell, Yaqi Xie, Woojun Kim, Katia Sycara, and Simon Stepputtis
    In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024
    IROS
  9. HiKER-SGG: Hierarchical Knowledge Enhanced Robust Scene Graph Generation
    Ce Zhang, Simon Stepputtis, Joseph Campbell, Katia Sycara, and Yaqi Xie
    In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024
    CVPR
  10. Embedding Symbolic Knowledge into Deep Networks
    Yaqi Xie, Ziwei Xu, Mohan S Kankanhalli, Kuldeep S Meel, and Harold Soh
    In Advances in Neural Information Processing Systems (NeurIPS), 2019
    NeurIPS
  11. Embedding Symbolic Temporal Knowledge into Deep Sequential Models
    Yaqi Xie, Fan Zhou, and Harold Soh
    In IEEE International Conference on Robotics and Automation (ICRA), 2021
    ICRA
  12. Multi-task trust transfer for human–robot interaction
    Harold Soh, Yaqi Xie, Min Chen, and David Hsu
    The International Journal of Robotics Research (IJRR), 2019
    IJRR
  13. Robot Capability and Intention in Trust-Based Decisions Across Tasks
    Yaqi Xie, Indu P Bodala, Desmond C. Ong, David Hsu, and Harold Soh
    In ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2019
    HRI
  14. Translating Natural Language to Planning Goals with Large-Language Models
    Yaqi Xie, Chen Yu, Tongyao Zhu, Jinbin Bai, Ze Gong, and Harold Soh
    In Preprint, 2023