Judy Hoffman

Computer Vision and Machine Learning Researcher


Upcoming Talks

Past Talks

  • Domain adaptation: From simulation data to real world training data
    Berkeley Deep Drive Symposium, November 2017
  • A General Framework for Domain Adversarial Learning
    OpenAI, July 2017
    Qualcomm Research, July 2017
    Berkeley Artificial Intelligence Research, April 2017
  • Invited Talk at Deep Learning Summit, San Francisco, January 2017.
  • Learning with Side Information through Modality Hallucination
    CVPR, June 2016
  • Cross-Modal Adaptation for RGB-D Detection
    ICRA, May 2016
  • Dissertation Talk April 2016
  • Adapting Deep Networks Across Domains, Modalities, and Tasks
    Stanford Computer Vision Seminar, January 2016
    Invited Talk at TASK-CV Workshop (ICCV), 2015. [slides pdf]
  • Simultaneous Transfer Across Domains and Tasks
    Bay Area Robotics Symposium, Berkeley, USA, 2015
  • Adapting Deep Networks to Real-World Problems
    Amazon Computer Vision PhD Symposium, October, 2015.
  • Large scale recognition through adaptation
    Berkeley-Stanford vision learning meeting, Berkeley, USA, 2015
  • Category Invariant Cross Modality Transfer
    Daghstuhl seminar Machine Learning with Interdependent and Non-identically Distributed Data, Schloss Dagstuhl, Germany, 2015.
  • Continuous Adaptation with Limited Target Labeled Data
    IST Austria Symposium on Computer Vision and Machine Learning, Vienna, Austria, 2015.
  • Transfer of Deep Vision (and Language) models for "TOT"
    DARPA meeting invited talk, 2014.
  • Large Scale Detection through Adaptation
    Baylearn (Presented Hoffman et al. NIPS 2014) [slides pptx] [slides pdf] [paper]
  • Efficient Learning of Domain Invariant Image Representations
    ICLR 2013. [slides pdf] [paper]
  • Discovering Latent Domains for Multisource Domain Adaptation
    WiML 2012. (Presentation of Hoffman et al. ECCV 2012) [slides pdf] [paper]