This workshop will consist of an opening remark, invited talks, contributed talks as well as a panel discussion session. The invited and contributed talks will be made available for offline viewing.
Invited Speakers
Location and Time
Location: Virtual
Date: July 18, 2020
Time: 11:30 AM - 2:00 PM EDT (EDT = Boston time)
Schedule
Schedule overview (EDT = Boston time):
- 11:30 AM - 11:35 PM EDT Introduction
- 11:35 AM - 1:05 PM EDT Invited talks
The invited speakers will be grouped into two sequential panels of 45 minutes each. Participants should watch the talks before joining the session, there will only be brief summaries by the speakers during the sessions. The abstracts and the videos of the invited talks can be found on the official ICML
website (accessible upon registration):
https://icml.cc/virtual/2020/workshop/5716 - for some of the talks, the link to the video is provided in the abstract.
The discussion sessions for the invited talks are organized as follows:
-
11:35 AM - 12:20 PM EDT Session 1
-
Aurélien Baillon, Follow the money, not the majority: Incentivizing and Aggregating Expert Opinions with Bayesian Markets
-
Yuqing Kong, Dominantly Truthful Multi-task Peer Prediction with a Constant Number of Tasks
-
Jens Witkowski, Incentive-Compatible Forecasting Competitions
-
12:20 AM - 1:05 PM EDT Session 2
-
Yiling Chen,
Strategic Considerations in Statistical Estimation and Learning
-
Ruoxi Jia, What is my data worth? Towards a Principled and Practical Approach for Data Valuation
-
Nihar Shah, Thwarting Dr. Deceit's Malicious Activities in Conference Peer Review
- 1:05 AM - 1:15 PM EDT Break
- 1:15 PM - 1:50 PM EDT Contributed talks
The contributed papers will be grouped into three, parallel panels of 35 minutes each. Participants should read papers and watch the talks before joining the session, there will only be brief summaries by the speakers during the sessions. The videos of the contributed talks can be found on the official ICML
website (accessible upon registration):
https://icml.cc/virtual/2020/workshop/5716 - for some of the talks, the link to the video is provided in the abstract.
The discussion sessions for the contributed talks are organized as follows:
-
1:15 PM - 1:50 PM EDT Session 1: Truthful Machine Learning
-
1:15 PM - 1:50 PM EDT Session 2: Learning and Strategies
-
1:15 PM - 1:50 PM EDT Session 3: Truthful Information Elicitation
Please refer to the list of accepted papers below.
- 1:50 AM - 2:00 PM EDT Concluding remarks
Accepted Papers
The following papers are accepted for presentation at the workshop:
-
Truthful Machine Learning
-
Learning and Strategies
- Jacob Abernethy, Bhuvesh Kumar, Thodoris Lykouris and Yinglun Xu. Bridging Truthfulness and Corruption-Robustness in Multi-Armed Bandit Mechanisms.
- Arnaud Fickinger, Simon Zhuang, Dylan Hadfield-Menell and Stuart Russell. Multi-Principal Assistance Games.
- Hanrui Zhang, Yu Cheng and Vincent Conitzer. Classification with Few Tests through Self-Selection.
- Daniel Ngo, Logan Stapleton, Nicole Immorlica, Vasilis Syrgkanis and Zhiwei Steven Wu. Incentivizing Bandit Exploration: Recommendations as Instruments.
- Yahav Bechavod, Katrina Ligett, Z. Steven Wu and Juba Ziani. Causal Feature Discovery through Strategic Modification.
-
Truthful Information Elicitation
- Jingyan Wang, Ivan Stelmakh, Nihar Shah and Yuting Wei. Debiasing Evaluations That are Biased by Evaluations.
- Nir Rosenfeld, Sophie Hilgard, Sai Srivatsa Ravindranath and David Parkes. From Predictions to Decisions: Using Lookahead Regularization.
- Steven Jecmen, Hanrui Zhang, Ryan Liu, Nihar Shah, Vincent Conitzer and Fei Fang. Mitigating Manipulation in Peer Review via Randomized Reviewer Assignments.
- Ivan Stelmakh, Nihar Shah and Aarti Singh. Catch Me if I Can: Detecting Strategic Behaviour in Peer Assessment.
- Ritesh Noothigattu, Nihar B. Shah and Ariel D. Procaccia. Loss Functions, Axioms, and Peer Review.