We study the core topics regarding trustworthy machine learning, with a focus on the statistical understructure of various topics under this umbrella topic, and also branching out the central understanding to solve multiple problems in topics such as robustness, causality, and interpretability, with a main application domain of vision.
We study computational biology because every progress we make has the potential to free millions from suffering.
We are devoted to developing methods that help understand the genetic basis of human complex traits. Our previous studies mostly focus on Alzheimer's disease and cancer.
We develop softwares for two purposes: 1) we always seek to deliver our innovations for domain experts to use free of any technical barriers; 2) we believe the trustworthiness of ML, by its definition, has users involved, thus incorporating users in the loop will strengthen the trustworthiness of methods. We are recruiting talents for Robustar now.
Updates |
|
---|---|
Aug. 2022 |
I'm starting my appointment as an assistant professor at iSchool at UIUC |
Aug. 2022 |
We presented our work on learning robust and invariant representations with data augmentation at KDD 2022. [Slides][Poster] |
Aug. 2022 |
We presented our work on a unified theme of robust machine learning titled toward learning human-aligned robust models at UAI 2022. [Poster] |
July 2022 |
We released the initial version of our software, Robustar, a GUI software that helps the user to indentify spurious features [Video][Code] |
July 2022 |
I gave an invited talk on Trustworthy AI-diagnosis of Alzheimer's Disease from MRI by Stanford University CNS lab [Slides] |
July 2022 |
I gave an invited talk on A Principled Unverstanding of Robust Machine Learning Methods by RIKEN Center for Advanced Intelligence Project [Video][Slides] |
June 2022 |
We presented our work on The Two Dimensions of Worst-case Training and the Integrated Effect for OOD Generalization at CVPR 2022. [Poster] |
May 2022 |
We presented our work on Gene Set Prioritization Guided by Regulatory Networks with p-values through KMM at RECOMB 2022. [Slides][Software] |
April 2022 |
I'm recognized as one of the top 50 AI+X rising young scholars by Baidu. Inc. |
Dec. 2021 |
I gave an invited talk on Toward Trustworthy Machine Learning to Understand the Personalized Genetic Basis of Alzheimer's Disease by Department of Bioinformatics at University of Pittsburgh [Slides] |
Dec. 2021 |
I defended my thesis on Toward Robust Machine Learning by Countering Superficial Features at LTI CMU [Thesis][Slides] |
We are working on the Robustar project, an interactive toolbox for pricise data annotation and robust vision learning.
Our goal is to allow domain experts to interact with the machine learning models through web applications to improve its robustness.
We have released the first version and are looking forward to meeting new collaborators at our at our repo. Check out our promotion video for more.
We are holding bi-weekly reading groups that unite the community with the same interest in trustworthy machine learning globally to learn and discuss recent papers.
The reading group usually convenes at noon on Thursday EST. The detailed schedule of the event and the summary of what we have read are announced through the twitter account @HaohanWang
We are exploring a new form of research collaborations by uniting the scholars of the trustworthy ML community globally for discussion and collaborations.
If you are interested in this collaboration, please fill in this form
Recent Talks |
|
---|---|
July 2022 |
Trustworthy AI-diagnosis of Alzheimer's Disease from MRI at Stanford University CNS lab [Slides] |
July 2022 |
A Principled Unverstanding of Robust Machine Learning Methods and Its Connections to Multiple Methods at RIKEN Center for Advanced Intelligence Project [Video][Slides] |
Dec. 2021 |
Toward Trustworthy Machine Learning to Understand the Personalized Genetic Basis of Alzheimer's Disease at Department of Bioinformatics at University of Pittsburgh [Slides] |
Mar. 2021 |
Robust Machine Learning with Emphasis on Countering Spurious Features at Data Science Initiative at Brown University |
Jan. 2021 |
High-frequency Component Helps Explain the Generalization of CNN at Aggregate Intellect [Video][Slides] |
Nov. 2020 |
Towards Trustworthy Machine Learning Inspired by High-frequency Data at Robotics Institute at Carnegie Mellon University [Slides] |
July. 2020 |
Toward Trustworthy Machine Learning for Scientific Discovery at Doctoral Symposium at ACM Conference on Health, Inference, and Learning |
April 2020 |
A Brief Overview of Trustworthy Machine Learning at Probabilistic Graphic Models lecture at Carnegie Mellon University [Slides] |
Feb. 2020 |
Learning Deconfounded Representations through Neural Networks, with Applications in Genetic Data at Center of Excellence for Computational Drug Abuse Research [Slides] |
Sept. 2019 |
Dealing with Confounding Factors in Deep Learning at Next Generation in Biomedicien Symposium at the Broad Institute [Slides] |
Sept. 2019 |
Deep Learning over Heterogeneous Data at Department of Bioinformatics at University of Pittsburgh [Slides] |
Haohan Wang (汪浩瀚) is an assistant professor in the School of Information Sciences at the University of Illinois Urbana-Champaign. His research focuses on the development of trustworthy machine learning methods for computational biology and healthcare applications, such as decoding the genomic language of Alzheimer's disease. In his work, he uses statistical analysis and deep learning methods, with an emphasis on data analysis using methods least influenced by spurious signals. Wang earned his PhD in computer science through the Language Technologies Institute of Carnegie Mellon University where he works with Professor Eric Xing. In 2019, Wang was recognized as the Next Generation in Biomedicine by the Broad Institute of MIT and Harvard because of his contributions in dealing with confounding factors with deep learning.