Hao Dong

haodong2.jpg

donghaospurs@gmail.com

I am an incoming Principal Investigator at ELLIS Institute Finland and Tenure-Track Assistant Professor at Tampere University. I received my Ph.D. from ETH Zurich, supervised by Prof. Eleni Chatzi and Prof. Olga Fink, Master’s degree from Aalto University, and Bachelor’s degree from Xi’an Jiaotong University.

My long-term research goal is to advance a new generation of intelligent AI systems that are robust, reliable, and adaptable. To operate safely in the wild, these systems must generalize across domain shifts (e.g., varying weather conditions or different sensor equipment), respond effectively to uncertain or unfamiliar inputs (such as out-of-distribution objects, misclassifications, and hallucinations), and continuously adapt to new domains and tasks (e.g., continual learning). Because the real world is inherently multimodal, exploiting complementary information across modalities is essential for strengthening overall system reliability. My core research areas include:

  • Multimodal Learning and Generative AI
  • Reliable and Trustworthy Machine Learning
  • Generalization, Adaptation, and Continual Learning of Foundation Models and Agentic Systems
  • Applications in Robotics, Healthcare, Industrial Monitoring, and Beyond

Hiring: I am always looking for motivated PhD students, postdocs, interns, and research collaborators. Please see the open positions page or reach out by email.

Short Bio

Hao Dong is an incoming Principal Investigator at ELLIS Institute Finland and Tenure-Track Assistant Professor at Tampere University. He received his Ph.D. from ETH Zurich, where he was advised by Prof. Eleni Chatzi and Prof. Olga Fink. He received his Master’s degree from Aalto University and Bachelor’s degree from Xi’an Jiaotong University. His research focuses on building robust, reliable, and adaptive intelligent machine learning systems, with particular emphasis on multimodal learning, domain generalization and adaptation, out-of-distribution detection, and their applications in robotics, healthcare, and industrial monitoring. He has published papers in top-tier AI conferences and journals, including NeurIPS, ICLR, CVPR, ECCV, ICRA, TPAMI, TNNLS, TMLR, and Nature Communications, with several papers selected as NeurIPS Spotlight and CVPR Highlight.

news

May 21, 2026 Excited to share that I will be joining ELLIS Institute Finland as Principal Investigator and Tampere University as Tenure-Track Assistant Professor.
Apr 17, 2026 Honored to be selected for the CVPR 2026 Doctoral Consortium and to receive a CVPR 2026 Travel Award.
Feb 18, 2026 Two papers accepted to CVPR 2026: “Adaptive Confidence Regularization for Multimodal Failure Detection” and “Towards Multimodal Domain Generalization with Few Labels (highlight)”.
Oct 01, 2025 Invited to participate in the Meta PhD Forum.
Sep 01, 2025 Paper accepted to NeurIPS 2025: “Extremely Simple Multimodal Outlier Synthesis for Out-of-Distribution Detection and Segmentation”.

selected publications

  1. PhD Thesis
    Towards Reliable Multimodal Learning for Real-world Systems
    Hao Dong
    ETH Zurich, 2026
  2. TPAMI
    Advances in Multimodal Adaptation and Generalization: From Traditional Approaches to Foundation Models
    Hao Dong, Moru Liu, Kaiyang Zhou, and 4 more authors
    IEEE Transactions on Pattern Analysis and Machine Intelligence, 2026
  3. ICLR
    Towards Robust Multimodal Open-set Test-time Adaptation via Adaptive Entropy-aware Optimization
    Hao Dong, Eleni Chatzi, and Olga Fink
    In International Conference on Learning Representations, 2025
  4. NeurIPS Spotlight
    MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities
    Hao Dong, Yue Zhao, Eleni Chatzi, and 1 more author
    In Advances in Neural Information Processing Systems, 2024
  5. ECCV
    Towards Multimodal Open-Set Domain Generalization and Adaptation through Self-supervision
    Hao Dong, Eleni Chatzi, and Olga Fink
    In European Conference on Computer Vision, 2024
  6. ICRA
    SuperFusion: Multilevel LiDAR-Camera Fusion for Long-Range HD Map Generation
    Hao Dong, Weihao Gu, Xianjing Zhang, and 5 more authors
    In International Conference on Robotics and Automation, 2024
  7. NeurIPS
    SimMMDG: A Simple and Effective Framework for Multi-modal Domain Generalization
    Hao Dong, Ismail Nejjar, Han Sun, and 2 more authors
    In Advances in Neural Information Processing Systems, 2023