Haichao Zhang 张海超
Senior Research Scientist
Horizon Robotics
hczhang1@gmail.com
I'm a Senior Research Scientist with Horizon Robotics, Cupertino.
Before that I was a Senior Research Scientist with Baidu Research, Sunnyvale (2016~2019) and Research Scientist with Amazon Go, Seattle (2014~2016).
Research Interests
reinforcement learning
sparse / adversarial learning
low / high level vision
image / video / language processing
learning-based planning and control
self-driving and autonomous robots
Awards and Honors
Best Student Paper Award 13th International Conference on Computer Vision (ICCV) 2011
Best Student Paper Honorable Mention 14th International Conference on Information Fusion (FUSION) 2011
Doctoral Consortium IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2013
Outstanding Doctoral Dissertation Award China Computer Federation (CCF) 2015
Updates
Policy Expansion Offline-to-Online RL (PEX) paper is accepted to ICLR'23
PaCo Multi-Task RL (NeurIPS'22) code is released on Github, Dec. 2022
Parameter Compositional Multi-Task Reinforcement Learning (PaCo) paper is accepted to NeurIPS'22
Generative Planning (ICLR'22) code is released on Github, April 2022
One paper on Generative Planning is accepted to ICLR'22 (Spotlight)
Serve as Area Chair for Winter Conference on Applications of Computer Vision WACV 2020
Bilateral Adversarial Training (ICCV'19) code is released on Github, Nov. 2019
Feature Scattering-based Adversarial Training (NeurIPS'19) code is released on Github, Oct. 2019
Feature Scattering-based Adversarial Training is accepted by NeurIPS'19
Two papers on Adversarial Robustness for Classification and Object Detection are accepted to ICCV'19
Towards Adversarially Robust Object Detection ICCV'19
Bilateral Adversarial Training ICCV'19
One paper on Visual Recognition under Adverse Conditions is published in IEEE TIP Sept. 2019
One paper on Improved Sequence-to-Sequence Learning is accepted to ICLR'19
One paper on Adversarial Text Generation is accepted by NeurIPS'18
One paper on Guided Feature Transformation for Embodied Agents is accepted to CoRL'18
One paper on Interactive Language Learning and One-shot Concept Learning is accepted to ACL'18
Present paper Interactive Grounded Language Acquisition and Generalization in a 2D World at ICLR'18
Present "Listen, Interact and Talk: Learning to Speak via Interaction" paper at ViGIL NIPS'17
Our XWorld simulator for Reinforcement Learning is now available on Github, Aug. 2017
Co-organize a tutorial on Low-Quality Visual Data Processing at CVPR'17
Receive Outstanding Doctoral Dissertation Award in 2015
One paper is accepted by CVPR'15
One paper on "Scale Adaptive Blind Deblurring" is accepted to NIPS'14
One paper on "Multi-Shot Imaging" is accepted to CVPR'14
Give an Oral presentation of the paper on "Non-Uniform Camera Shake Removal" at NIPS'13 Oral
One paper on Coupled Adaptive Sparse Prior based Blind Deblurring is accepted by CVPR'13 and CVPR Doctoral Consortium
Two papers on Joint Restoration and Recognition and Joint Dynamic Sparse Representation are accepted by ICCV'11
One paper on Non-Local Kernel Regression for unified local and non-local image modeling is accepted by ECCV'10