NeurIPS 2023 Reveals Outstanding Paper Awards Across Record Submissions
NeurIPS 2023 officially announces its award-winning papers, highlighting top research across main and dataset tracks from a record-breaking pool of over 13,000 submissions. Notable winners include studies on efficient privacy auditing and the debated emergent abilities of large language models.
NeurIPS 2023 officially unveils its prestigious paper awards, recognizing top achievements in machine learning research. The conference organizers evaluate a record-breaking 13,300 submissions with the help of 968 area chairs, 98 senior area chairs, and 396 ethics reviewers, ultimately accepting 3,540 papers into the program.
The awards feature two Outstanding Main Track Papers, two Main Track Runner-Ups, and two Outstanding Datasets and Benchmark Track Papers. One notable main track winner introduces a highly efficient scheme for auditing differentially private machine learning systems using just a single training run, drastically reducing the computational cost compared to standard methods.
Another standout winning paper questions whether the emergent abilities of large language models are actually a mirage rather than a real phenomenon. Attendees view these award-winning presentations at designated poster and oral sessions throughout the conference, while the organizers express their gratitude to the dedicated awards committee members who navigate conflict-of-interest guidelines to select the honorees.