Zhe Gan

I am a Senior Researcher at Microsoft Dynamics 365 AI Research, primarily working on Vision-and-Language Representation Learning, Generative Pre-training, and Adversarial Machine Learning. I also have broad interests on other machine learning topics. I received my Ph.D. degree from Duke University in Spring 2018. Before that, I received my Master's and B.Sc. from Peking University in 2013 and 2010, respectively. My Ph.D. advisor is Lawrence Carin. I can be reached at zhe.gan@microsoft.com.

I am serving as an Area Chair for ICLR 2021 and NeurIPS 2020, and an Interactive Arts Chair at ACM Multimedia 2020. I received AAAI-20 Outstanding SPC Award. I served as an Area Chair for NeurIPS 2019, and a Senior Program Committee (SPC) member for AAAI 2020.

Research Highlights:

[2020/07] Will serve as an Area Chair for ICLR 2021.

[2020/07] Two papers got accepted to ECCV 2020. (1) UNITER: a state-of-the-art pre-trained Vision+Language (V+L) model; (2) VALUE (ECCV Spotlight): the first work on probing pre-trained V+L models.

[2020/06] Please check out our new VILLA paper, the first known effort that studies large-scale adversarial training for vision-and-langauge representation learning in both pre-training and finetuning stages.

[2020/06] Two papers got accepted to ICML 2020. (i) CLUB: a novel upper bound of mutual information that is deeply connected with contrastive learning. (ii) GOT: a graph optimal transport framework for cross-domain alignment that can be used for V+L and NLP problems, such as VQA and NMT.

[2020/06] At this year's CVPR, we will give a tutorial on "Recent Advances in Vision-and-Language Research", covering the recent popular multi-modal pre-training methods and other topics. More details are provided in the website here.

[2020/04] Two CVPR and three ACL papers got accepted, respectively. CVPR papers cover topics including: (i) high-resolution image synthesis from salient object layout, and (ii) a new dataset for video-and-language understanding. ACL papers cover topics include: (i) text summarization based on discourse units, (ii) BERT for text generation, and (iii) text generation that models the distant future.

[2020/03] Will serve as an Area Chair for NeurIPS 2020.

[2020/01] I received AAAI-20 Outstanding SPC Award.

[2019/09] Our new work UNITER achieves SOTA on 6 Vision-and-Language tasks across 9 datasets (VQA, VCR, NLVR, Img-Txt Retrieval, Visual Entailment, Referring Expression).

[2019/09] Our latest Adversarial Training model has beaten Facebook's RoBERTa on GLUE benchmark. Paper is available here.

[2019/08] 4 papers got accepted to EMNLP. Topics include (i) BERT model compression, (ii) domain adaptation for MRC, (iii) domain adaptation for text style transfer, and (iv) image caption evaluation.

[2019/07] One paper got accepted to ICCV, using relation-aware graph attention for VQA.

[2018/09] Two papers got accepted to NIPS 2018. The first one proposes Adversarial Information Maximization (AIM) for conversational response generation. The second one proposes feature mover GAN for neural text generation. The acceptance ratio this year is 1011/4856=20.82%.

[2018/02/20] PhD thesis defended.

[2018/02] One paper accepted to CVPR 2018. GAN plus attention results in our AttnGAN, generates realistic images on birds and COCO datasets. See the news coverage here: Drawing AI, among other places.

[2017/09] 4 papers got accepted to NIPS 2017; three of them are on deep generative models, including VAE and GAN variants; the other one is on deconvolutional paragraph representation learning. The acceptance ratio of NIPS 2017 is 678/3240= 20.93%.