Jiashu Xu 徐家澍

Research Scientist at NVIDIA

profile.jpg

Hi there 👋

I am a research scientist at NVIDIA. I was a master student at Harvard University studying computer science. Previously I double majored in applied mathematics and computer science at USC. I had the honor to work with Prof. Hima Lakkaraju at Harvard SEAS, and Prof. Muhao Chen at USC. I have spent a bit of time as research intern at Amazon Science team working on LLM and as student collaborator at Microsoft Research working on synthetic data generation.

My current research interests is in Reliable AI. Particularly,

  • AI safety [1, 2, 3]
  • Training AI that excels in low-resource regime, through indirect supervision [1, 2] or synthetic data [3, 4, 5, 6]
  • Explanation and how we can learn from explanation [1, 2, 3]

This is my girlfriend 😍Carmen and me

Selected Publications

  1. fingerprintv2.png
    NAACL Oral
    Instructional Fingerprinting of Large Language Models
    In Proceedings of the Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL) , 2024
  2. dreamdistribution.png
    ECCV Under Review
    DreamDistribution: Prompt Distribution Learning for Text-to-Image Diffusion Models
    In review of The European Conference on Computer Vision (ECCV), 2024
  3. BVS.png
    CVPR Highlight
    BEHAVIOR Vision Suite: Customizable Dataset Generation via Simulation
    In Conference on Computer Vision and Pattern Recognition (CVPR) , 2024
  4. can-nli.jpg
    ACL Oral
    Can NLI Provide Proper Indirect Supervision for Low-resource Biomedical Relation Extraction?
    Jiashu XuMingyu Derek Ma, and Muhao Chen
    In Association for Computational Linguistics (ACL) , Jul 2023
  5. Instructions as Backdoors: Backdoor Vulnerabilities of Instruction Tuning for Large Language Models
    In Proceedings of the Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL) , Jul 2024
  6. beyond-generation.png
    Arxiv Extension
    Beyond Generation: Harnessing Text to Image Models for Object Detection and Segmentation
    arXiv preprint arXiv:2309.05956, Jul 2023
  7. X-Norm: Exchanging Normalization Parameters for Bimodal Fusion
    Yufeng Yin*Jiashu Xu*Tianxin Zu, and Mohammad Soleymani
    In Proceedings of the 2022 International Conference on Multimodal Interaction (ICMI) , Jul 2022
  8. Neural-Sim: Learning to Generate Training Data with NeRF
    In European Conference on Computer Vision (ECCV) , Jul 2022