I am a 4-th year PhD student at Stanford Computer Science advised by Stefano Ermon. Before coming to stanford, I completed my undergrad at Tsinghua University.

Contact: sjzhao at stanford dot edu

Research Interests

(Ranked by the amount of time I spend thinking about each topic, as of Jan 2020)

• Variational definitions of information and machine learning quantities
• Individual risk, Operational interpretations for probability and forecast, Calibration
• Deep generative models, high dimensional distribution modeling, representation learning
• Active learning and online learning
• Adversarial examples
• Machine learning applications of information theory

Publications

Recent Work

• A Information Theoretic Framework for Adversarial Defense and Attack

• Individual Calibration by Randomized Forecast

Conference

• A framework for Sample Efficient Interval Estimation with Control Variates
  Shengjia Zhao, Christopher Yeh, Stefano Ermon (AISTATS’2020)

• Permutation Invariant Graph Generation via Score-Based Generative Modeling
  Chenhao Niu, Yang Song, Jiaming Song, Shengjia Zhao , Aditya Grover, Stefano Ermon (AISTATS’2020)

• A Theory of Usable Information under Computational Constraints
  Yilun Xu, Shengjia Zhao , Jiaming Song, Russell Stewart, Stefano Ermon (ICLR’2020 Oral)

• Adaptive Antithetic Sampling for Variance Reduction
  Hongyu Ren*, Shengjia Zhao* , Stefano Ermon (ICML’2019)

• Learning Neural PDE Solvers with Convergence Guarantees
  Jun-Ting Hsieh*, Shengjia Zhao*, Lucia Mirabella, Stefano Ermon (ICLR’2019)

• Regular LDPC Construction for Sparse Hashing
  Jonathan Kuck, Tri Dao, Shengjia Zhao, Burak Burtan, Stefano Ermon (UAI’2019)

• Learning Controllable Fair Representations
  Jiaming Song, Pratyusha Kalluri, Aditya Grover, Shengjia Zhao, Stefano Ermon (AISTATS’2019)

• InfoVAE: Balancing Learning and Inference in Variational Autoencoders
  Shengjia Zhao, Jiaming Song, Stefano Ermon [arXiv] (AAAI’2019)

• Bias and Generalization in Deep Generative Models: An Empirical Study
  Shengjia Zhao*, Hongyu Ren*, Arianna Yuan, Jiaming Song, Noah Goodman, Stefano Ermon (NeurIPS’2018 Spotlight )

• Amortized Inference Regularization
  Rui Shu, Hung H Bai, Shengjia Zhao, Stefano Ermon (NeurIPS’2018)

• A Lagrangian Perspective on Latent Variable Generative Models
  Shengjia Zhao, Jiaming Song, Stefano Ermon (UAI’2018 Oral) [arXiv]

• A-NICE-MC: Adversarial Training for MCMC
  Jiaming Song, Shengjia Zhao, Stefano Ermon (NeurIPS’2017) [arXiv]

• Learning Hierarchical Features from Generative Models
  Shengjia Zhao, Jiaming Song, Stefano Ermon (ICML’2017) [arXiv] [code]

• Adaptive Concentration Inequalities for Sequential Decision Problems
  Shengjia Zhao, Enze Zhou, Ashish Sabharwal, Stefano Ermon (NeurIPS’2016) [pdf]

• Closing the Gap Between Short and Long XORs for Model Counting
  Shengjia Zhao, Sorathan Chaturapruek, Ashish Sabharwal, Stefano Ermon (AAAI’2016) [arXiv]

Miscellaneous

Awards and Fellowships

• JP Morgan PhD Fellowship (2019)
• Qualcomm Innovation Fellowship (QInF) (2018)
• Qualcomm Scholarship (2016)
• Google Excellence Scholarship (2015)

Professional Service

• Reviewer: NeurIPS (2017, 2019), ICLR (2019, 2020), ICML (2019, 2020)
• Organizer: Information Theory and Machine Learning (ITML) Workshop (NeurIPS’2019)
• Teaching: CS228 (Head TA)