Mike Wu
me[at]mikehwu.com
Google Scholar | Github
 LinkedIn | @mike_h_wu

I received my PhD in June '22 from Stanford. I was advised by Noah Goodman. My research is in generative models and self-supervised learning, with applications to education.


efore, I worked as a research engineer in the applied ML team at Meta Research, and as a SWE at an AI startup called Lattice Data. Prior to working, I worked with Frank Wood at Oxford University on probabilistic programming and with Finale Doshi-Velez on interpretable ML for healthcare.


Mike Wu
me[at]mikehwu.com
Google Scholar | Github
 LinkedIn | @mike_h_wu

I received my PhD in June '22 from Stanford. I was advised by Noah Goodman. My research is in generative models and self-supervised learning, with applications to education.


Before, I worked as a research engineer in the applied ML team at Meta Research, and as a SWE at an AI startup called Lattice Data. Prior to working, I worked with Frank Wood at Oxford University on probabilistic programming and with Finale Doshi-Velez on interpretable ML for healthcare.

News

  • Hot off the press! New paper on approximate probabilistic inference as masked language modeling. Check out transformers being Bayesian.

Thesis

  • Extensions and Applications of Deep Probabilistic Inference for Generative Models.
    Mike Wu.
    (pdf) (code)

Publications

  • Human-Level Rubric-Grading of Novel, Free-Response Exams Through Meta-Learning
    Mike Wu, Alan Cheng, Noah Goodman, Chelsea Finn, Chris Piech.
    Under Review, PNAS 2022.
  • Foundation Posteriors for Approximate Probabilistic Inference.
    Mike Wu, Noah Goodman.
    NeurIPS, 2022.
    (pdf) (poster)
  • Modeling Item Response Theory with Stochastic Variational Inference.
    Mike Wu, Richard L. Davis, Benjamin W. Domingue, Chris Piech, Noah Goodman.
    ArXiv.
    (pdf)
  • Differentiating small-scale subhalo distributions in CDM and WDM models using persistent homology.
    Jessi Cisewski-Kehe, Brittany Terese Fasy, Wojciech Hellwing, Mark R. Lovell, Pawel Drozda, Mike Wu.
    Physical Review D, 2022.
    (pdf) (journal)
  • Know Thy Student: Interactive Learning through Gaussian Processes.
    Rose Wang, Mike Wu, Noah Goodman.
    ICLR Workshop on Collective Learning Across Scales, 2022.
    (pdf)
  • Temperature as Uncertainty in Contrastive Learning.
    Oliver Zhang, Mike Wu, Jasmine Bayrooti, Noah Goodman.
    NeurIPS Workshop on Self-Supervised Learning, 2021.
    (pdf) (poster) (code)
  • Tradeoffs Between Contrastive and Supervised Learning: An Empirical Study.
    Ananya Karthik, Mike Wu, Noah Goodman, Alex Tamkin.
    NeurIPS Workshop on Self-Supervised Learning, 2021.
    (pdf) (poster)
  • Improving Compositionality of Neural Networks by Decoding Representations to Inputs.
    Mike Wu, Noah Goodman, Stefano Ermon.
    NeurIPS, 2021.
    (pdf) (talk)
  • HarperValleyBank: A Domain-Specific Spoken Dialog Corpus.
    Mike Wu, Jonathan Nafziger, Anthony Scodary, Andrew Maas.
    Workshop on Machine Learning in Speech and Language Processing (MLSLP), 2021.
    Used for Stanford CS224S: Spoken Language Processing.
    (pdf) (code) (poster) (cs224s)
  • Optimizing for Interpretability in Deep Neural Networks with Simulable Decision Trees.
    Mike Wu, Sonali Parbhoo, Michael C. Hughes, Volker Roth, Finale Doshi-Velez.
    Journal of Artificial Intelligence Research (JAIR) Vol. 72, 2021.
    (pdf) (journal) (CHAI talk)
  • Generative Grading: Near Human-level Accuracy for Automated Feedback on Richly Structured Problems.
    Ali Malik (*), Mike Wu (*), Vrinda Vasavada, Jinpeng Song, John Mitchell, Noah Goodman, Chris Piech.
    Educational Data Mining (EDM), 2021.
    (pdf)
  • A Simple Framework for Uncertainty in Contrastive Learning.
    Mike Wu, Noah Goodman.
    ArXiv, 2020.
    (pdf)
  • On Mutual Information in Contrastive Learning for Visual Representations.
    Mike Wu, Chengxu Zhuang, Milan Mosse, Daniel Yamins, Noah Goodman.
    ArXiv, 2020.
    (v1 pdf) (v0 pdf) (talk)
  • Variational Item Response Theory: Fast, Accurate, and Expressive.
    Mike Wu, Richard L. Davis, Benjamin W. Domingue, Chris Piech, Noah Goodman.
    Educational Data Mining (EDM), 2020.
    (pdf) (code) (video) (oral) (best paper award)
  • Regional Tree Regularization for Interpretability in Black Box Models.
    Mike Wu, Sonali Parbhoo, Michael C. Hughes, Ryan Kindle, Leo Celi, Maurizio Zazzi, Volker Roth, Finale Doshi-Velez.
    AAAI, 2020.
    (pdf) (oral) (poster) (AAAI talk)
  • Multimodal Generative Models for Compositional Representation Learning.
    Mike Wu, Noah Goodman.
    ArXiv, 2019.
    (pdf)
  • Meta-Amortized Variational Inference and Learning.
    Mike Wu (*), Kristy Choi (*), Noah Goodman, Stefano Ermon.
    AAAI, 2020.
    NeurIPS Workshop on Bayesian Deep Learning, 2019.
    (pdf) (workshop oral) (code)
  • Pragmatic inference and Visual Abstraction Enable Contextual Flexibility during Visual Communication.
    Judith Fan, Robert X.D. Hawkins, Mike Wu, Noah Goodman.
    Computational Brain & Behavior (COBB) Vol. 3, 2019.
    (pdf) (code)
  • Differentiable Antithetic Sampling for Variance Reduction in Stochastic Variational Inference.
    Mike Wu, Noah Goodman, Stefano Ermon.
    AISTATS, 2019.
    (pdf) (code) (poster)
  • Zero Shot Learning for CodeEducation: Rubric Sampling with Deep Learning Inference.
    Mike Wu, Milan Mosse, Noah Goodman, Chris Piech.
    AAAI, 2019.
    (pdf) (oral) (outstanding student paper award) (code) (poster)
  • Multimodal Generative Models for Scalable Weakly Supervised Learning.
    Mike Wu, Noah Goodman.
    NeurIPS, 2018.
    (pdf) (code)
  • Spreadsheet Probabilistic Programming.
    Mike Wu, Yura Perov, Frank Wood, Hongseok Yang, William Smith.
    PROBPROG, 2018.
    (pdf)
  • Beyond Sparsity: Tree Regularization of Deep Models for Interpretability.
    Mike Wu, Michael C. Hughes, Sonali Parbhoo, Maurizio Zazzi, Volker Roth, Finale Doshi-Velez.
    AAAI, 2017.
    NeurIPS Workshop on Transparency in Machine Learning (TiML), 2017.
    (pdf) (code) (oral) (workshop oral) (TiML talk) (AAAI talk)
  • Predicting Intervention Onset in the ICU with Switching Statespace Models.
    Marzyeh Ghassemi, Mike Wu, Michael C. Hughes, Finale Doshi-Velez.
    AMIA Clinical Research Informatics (CRI), 2017.
    (pdf) (oral) (best paper finalist)
  • Understanding Vassopressor Intervention and Weaning: Risk Prediction in a Public Heterogeneous Clinical Time Series Database.
    Mike Wu, Marzyeh Ghassemi, Mengling Feng, Leo Anthony Celi, Peter Szolovitz, Finale Doshi-Velez.
    Journal of the American Medical Informatics Association (JAMIA), 2016.
    (pdf) (journal)
  • Edge-based Crowd Detection from Single Image Datasets.
    Mike Wu, Madhu Krishnan.
    IJCSI, 2013.
    (pdf)
  • Autonomous Mapping and Navigation through Utilization of Edge-based Optical Flow and Time-to-Collision.
     Madhu Krishnan, Mike Wu, Young Kang, Sarah H. Lee.
    Journal of Engineering and Applied Sciences (ARPN), 2012.
    ISEF SemiFinalist, 2012.
    (pdf) (journal)

Conference Abstracts

  • Gravitational Contrastive Learning: Inducing Structure During Instance Discrimination.
    Jasmine Bayrooti, Mike Wu, Alex Tamkin, Noah Goodman.
    Stanford CURIS, 2020.
    (pdf) (outstanding poster award)
  • Modeling Contextual Flexibility in Visual Communication.
    Judith Fan, Robert X.D. Hawkins, Mike Wu, Noah Goodman.
    Vision Sciences Society (VSS), 2018.
  • Investigating the Cosmic Web with Topological Data Analysis.
    Jessi Cisewski-Kehe, Mike Wu, Brittany Fasy, Wojciech Hellwing, Mark Lovell, Alessandro Rinaldo, Larry Wasserman.
    American Astronomical Society (AAS), 2018.
    Yale Undergraduate Thesis.
  • Position and Vector Detection of Blind Spot motion with the Horn-Schunck Optical Flow.
    Stephen Yu, Mike Wu. XSEDE, 2011.
    Siemens Competition Semifinalist, 2012.
    Intel International Science Fair (ISEF) Finalist, 2011: 3rd Place.
    (best student poster)