RobustML Workshop
Workshop at ICLR 2021
May 7 (Fri) 8am - 5pm (Pacific Time)
RobustML is an online workshop on robust and reliable Machine Learning in the real world, at the ICLR 2021 conference. It is on May 7th (Fri) 8am - 5pm (Pacific Time, GMT-7).
Official link: https://iclr.cc/virtual/2021/workshop/2129
A 1-day workshop of talks, forums, and poster sessions
2 Poster Sessions
Gather.Town Link
5 Invited Talks & 1 Forum
30-min each talk + 15-min QA
5 Oral Presentations
Presentation of selected papers
Schedule (Pacific time, GMT-7)
8:05 - Opening remarks
Opening remarks by Aditi Raghunathan (Stanford)
8:15 - Oral Presentation 1
Contributed talk: "On the Benefits of Defining Vicinal Distributions in Latent Space"
Puneet Mangla, Vedant Singh, Shreyas Havaldar, Vineeth Balasubramanian
8:30 - Talk 1: Ece Kamar
Microsoft (Senior principle research area manager)
8:30 - Invited talk: Ece Kamar
Ece Kamar is a Senior Principal Research Area Manager at Microsoft Research. She focuses on human-centered AI and manage the Adaptive Systems and Interaction Group. She earned her Ph.D. in computer science from Harvard University advised by Prof. Barbara Grosz.
9:00 - Live Q&A: Ece Kamar
Moderator: Aditi Raghunathan (Stanford)
9:15 - Talk 2: Kendra Albert
Harvard Law School (Clinical instructor & lawyer)
9:15 - Invited talk: Kendra Albert
Kendra Albert is a Clinical Instructor at the Cyberlaw Clinic and was formerly an associate at Zeitgeist Law PC, a boutique technology law firm in San Francisco.
9:45 - Live Q&A: Kendra Albert
Moderator: Aditi Raghunathan (Stanford)
10:00 - Poster session 1
Gather Town link
11:00 - Oral Presentation 2
Contributed talk: "Neural Lower Bounds for Verification"
Florian Jaeckle, M. Pawan Kumar
11:15 - Talk 3: Percy Liang
Stanford University (Associate Professor)
11:15 - Invited talk: Percy Liang
Percy Liang is an associate professor at Stanford University. He focuses on developing trustworthy systems that can communicate effectively with people and improve over time through interaction.
11:45 - Live Q&A: Percy Liang
Moderator: Zhijing Jin (Max Planck Institute)
12:00 - Panel discussion
Panel discussion: "Real world challenges for robustness in machine learning"
Moderator: Zhijing Jin (Max Planck Institute)
Panelist: Kendra Albert, Finale Doshi-Velez, Ece Kamar, Nicolas Papernot
Kendra Albert
Harvard
Finale Doshi-Velez
Harvard
Ece Kamar
Microsoft
Nicolas Papernot
UToronto
13:00 - Break
14:00 - Oral Presentation 3
Contributed talk: "Pervasive Label Errors in Test Sets Destabilize Machine Learning Benchmarks"
Curtis G Northcutt, Anish Athalye, Jonas Mueller
14:15 - Talk 4: Ruoxi Jia
Virginia Tech
Ruoxi Jia is a professor at Virginia Tech. Her research interests lie broadly in the span of machine learning, security, privacy, and cyber-physical systems.
Moderator: Eric Wong (MIT)
14:45 - Oral Presentation 4
Contributed talk: "A Causal Lens for Controllable Text Generation"
15:00 - Poster session 2
Gather Town link
16:00 - Oral Presentation 5
Contributed talk: "On Calibration and Out-of-Domain Generalization"
Yoav Wald, Amir Feder, Daniel Greenfeld, Uri Shalit
16:15 - Talk 5: Bo Li
UIUC
16:15 - Invited talk: Bo Li
Bo Li is an assistant professor at University of Illinois at Urbana-Champaign. Her research focuses on machine learning, security, privacy, and game theory, including exploring vulnerabilities of machine learning systems to various adversarial attacks, and endeavors to develop real-world robust learning systems.
16:45 - Live Q&A: Bo Li
Moderator: Eric Wong (MIT)
17:00 - Closing remarks
Closing remarks by Eric Wong (MIT)
Organizers
Amazon (Applied Scientist)
MIT (postdoc)
Microsoft (Principal Researcher)
UNC Chapel Hill (Associate Professor)
Technion (Assistant Professor)
UCLA (Assistant Professor)
University of Virginia (Associate Professor)
PhD (Max Planck Institute)
PhD (Stanford)
UNC Chapel Hill (PhD)