We invite researchers and industry professionals to submit their papers on negative results, failed experiments, and unexpected challenges encountered in applying deep learning to real-world problems across industry and science. The primary goal of this workshop is to create a platform for open and honest discussion about the hurdles and roadblocks in applying deep learning. We believe that sharing these experiences is crucial for the advancement of the field, providing valuable insights that can prevent others from repeating the same mistakes and fostering a culture of transparency and learning. We invite submissions from novel, ongoing, and unpublished research that apply deep learning to various domains including, but not limited to, social sciences, biology, physics, chemistry, engineering, robotics, psychology, healthcare, neuroscience, marketing, economics, or finance. Submitted papers should contain the following four elements:
- A use case that was tackled with deep learning.
- A solution for this type of use case was proposed in the deep learning literature
- A description of the (negative) outcome in the solution.
- An investigation (and ideally an answer) to the question of why it did not work as promised by the deep learning literature.
The potential reasons for failure may include but are not limited to data-related issues (e.g., distribution shift, bias, label quality, noisy measurement, quality of simulated data), model limitations (e.g., assumption violations, robustness, interpretability, scalability, representation misalignment), and deployment challenges (e.g., computational demands, hardware constraints). Besides these four points, papers will be assessed on:
- Rigor and transparency in the scientific methodologies employed.
- Novelty and significance of insights.
- Quality of discussion of limitations.
- Reproducibility of results.
- Clarity of writing.
Accepted papers will be made available on Open Review and the workshop website. Selected papers with exemplary scientific rigor, insightful findings, and excellent presentation will be nominated by reviewers for optional inclusion in a special issue of PMLR. Authors of selected papers can opt out of formal publication in PMLR. Proceedings of our last workshop can be found here. Submissions that do not appear in the proceedings will be non-archival, and authors are free to submit them to other venues.
Furthermore, the program chairs will nominate papers for the spotlight and contributed talks as well as two awards: (1) “Entropic Award” for the most surprising negative result and (2) “Didactic Award” for the most well-explained and pedagogical papers.
This year, ICLR is discontinuing the separate “Tiny Papers” track, and is instead requiring each workshop to accept short (2 pages in ICLR format) paper submissions, with an eye towards inclusion; see https://iclr.cc/Conferences/2025/CallForTinyPapers for more details. Authors of these papers will be earmarked for potential funding from ICLR, but need to submit a separate application for Financial Assistance that evaluates their eligibility. This application for Financial Assistance to attend ICLR 2025 will become available on https://iclr.cc/Conferences/2025/ at the beginning of February and close on March 2nd.