What’s new for the Position Paper Track at NeurIPS 2026

March 30 2026 What’s new for the Position Paper Track at NeurIPS 2026 Communication Chairs 2026 2026 Conference We are excited to announce the Position Paper Track is back at NeurIPS 2026 for the second year. Last year was a crucial experiment that created a forum for discussion on hot topics in our field. As a newer track at NeurIPS, we made several experimental decisions last year intended to promote discussion and improve positive interactions during review. We have been carefully listening to the community’s feedback, both positive and critical, and are excited to announce in this blog post new directions for the track this year. Aligning timeline and standards, broadening scope, and providing clearer definitions of rigor As we learn more about community expectations, we continue to evolve the track accordingly, and this year we are seeking to better align our paper acceptance process and timelines for submission, review, and notification with the other conference tracks. Working towards standardised practices for acceptance decisions wherever feasible (informed by post-conference reports from last year’s Main and Datasets and Benchmarks chairs) will ensure more consistent standards for quality and  acceptance rates across NeurIPS tracks. We also heard feedback from the community that the lag in review and decision timeline relative to other tracks was disappointing, as it caused authors to miss deadlines for re-submission to other venues. We are therefore planning to adhere closely to timelines adopted by the other tracks this year. We’re also aiming to broaden the scope of papers accepted to the Position Paper Track. Many community members perceived a bias against papers with technical topics in 2025. While part of this balance occurred because some papers submitted to the Position Paper Track were better suited for the Main Program (we desk rejected many papers that were effectively “our method is superior and the community should adopt or investigate it”), we recognize that it may also reflect a reviewing pool skewed towards certain kinds of expertise. To address this, we’re coordinating with the other tracks to recruit reviewers from a shared pool. Not only will this broaden the range of expertise for reviewers on the Position Paper Track, but it will also reduce burden on reviewers by ensuring that they do not get assigned to a full review load for multiple tracks simultaneously. We hope this broadening of reviewer expertise will better equip us to address both topics that are of more widespread interest for the entire community, and topics that may be specific to some subcommunities. We ask the community to please submit papers engaging with a wide range of topics . We welcome everything from more technical topics like “Bayesian learning is crucial for LLMs”, to topics with more widespread accessibility like “Deployment of agentic AI is hurting society”, to topics for specific application areas like “Virtual cell models must predict more than gene expression” – and obviously, this list is not exhaustive. We look forward to your perspectives! Finally, if you go to our Call for Papers, you’ll notice that we’ve substantially updated our text this year. Our aim was to better define what high-quality, rigorous submissions look like for the Position Paper Track, and we encourage all authors to carefully read our new criteria. We are putting greater emphasis on how well-argued the position is, not just the likelihood that a position will create discussion in the community. We make it clearer that position papers should support their positions with evidence. We also ask that authors make a case that topics are interesting for contemporary discussion in their papers, not just assume that the topic is widely held to be “hot”. Because position papers are relatively new for machine learning conferences, standards for quality and track boundaries are still forming. To help refine this, we are asking authors to prepare a short statement explaining why their submission is best suited to the position paper track. This statement will be used to assess that submissions are indeed aligned with the criterion that we are putting out this year, but also to refine criterion for future years. Improving author-reviewer interaction and review Last year, we experimented with review processes, moving rebuttals to an author survey and an adjudication process. While we noticed some positive dynamics from both and the adjudication process, although noisy, was instrumental in identifying some cases of misconduct in the review process, we also heard feedback from the authors that they were frustrated at not being able to discuss opportunities for changes with the reviewers. Hence, this year, recognizing that interaction between authors and reviewers is important, we intend to align with the main conference in creating a period for author-reviewer discussion. In addition, the Position Paper Track will be participating in the Responsible Reviewing initiative with the Main Program and Evaluation and Datasets track. Last year, around 35% of reviewers were not responsive in their reviews. This created a huge need for emergency reviewers and delayed our timeline in releasing decisions. As last year was our first year in hosting this track, we were unprepared for the high rate of missing reviews. We are regretful this happened as it meant some authors missed deadlines to resubmit rejected papers to other venues. Learning from experience, this year we intend to adopt the same policies by other tracks at the conference to safeguard reviewing quality and timeline. Crucially, the Responsible Reviewing initiative will apply across tracks: for example, reviewers who do not submit reviews on time for the Position Paper Track can expect to have reviews for their papers withheld even if those papers are under the Main Program or Evaluation and Datasets track. Please provide feedback! The Position Paper Track still remains a new track for NeurIPS, so the goal of releasing this blog post is to declare our intentions this year on how we will refine and improve it. However, we remain committed to learning throughout the process this year, and while some decisions may be fixed early on, others may have flexibility as conference planning proceeds. For this reason, we are especially keen on receiving feedback from the community. Do you think our directions this year are going the right way? Do you disagree? Do you have ideas? Please reach out by emailing us at position-track-chairs@neurips.cc!