The Best Paper Runner-Up award at NeurIPS was given to work coauthored by Shay Moran, a faculty member in our department as well as in the Faculty of Computer Science and the Faculty of Data and Decision Sciences, highlighting a breakthrough on a central open problem in online learning. This year, more than 21,000 papers were submitted to NeurIPS, of which approximately 6,000 were accepted, and only seven papers were selected for outstanding recognition – including this one.
The paper, “Optimal Mistake Bounds for Transductive Online Learning,” presents a solution to a significant open problem that has been studied for three decades.
The research team includes Zachary Chase, who completed a postdoctoral fellowship in our faculty, Jonathan Shafer, formerly a researcher in our department, and Steve Hanneke of Purdue University. Their work provides a precise quantification of the gap between standard online learning and the transductive setting, demonstrating tight mistake bounds that underscore the role of unlabeled data in online learning, in contrast to its behavior in the PAC framework.
Recognition as a Best Paper Runner-Up at NeurIPS reflects the theoretical significance of the work and its contribution to the broader understanding of computational learning.
We extend our congratulations to Shay and his collaborators.
Official NeurIPS announcement:
https://lnkd.in/gziShEec
