Hi all, We’re excited to announce the Call for Papers <https://unireps.org/2026/call-for-papers> for the 4th edition of the UniReps Workshop, happening in Paris <https://unireps.org/2026/>! The workshop is tentatively planned for mid-December, with a target date between December 11 and December 15, 2026, coinciding with NeurIPS. The exact date will be confirmed once the venue arrangements are finalized. You can indicate your preferred date using this form <https://forms.gle/nfpXEMiTWfoyKiX87>. Our workshop theme focuses on a central question: Why, when, and how do distinct learning processes yield similar representations, and to what extent can these be unified? Neural systems, whether artificial or biological, often converge on structurally similar internal representations when exposed to semantically related inputs. This workshop brings together researchers from Neuroscience, Artificial Intelligence, and Cognitive Science to build a unified understanding of this phenomenon and explore its theoretical and practical implications. We invite contributions addressing the following questions: - When? Under what conditions do similar representations emerge, and how can we measure this? - Why? What mechanisms, biological or computational, drive representational convergence? - What for? How can we harness this understanding for modular deep learning, model merging, transfer learning, and beyond? Topics of Interest Here’s a non-exhaustive list of themes we’re excited to explore, targeting especially their relationship with the questions listed above: - Model merging, stitching, and compositionality - Representational alignment and similarity measures - Identifiability and symmetry in neural models - Learning dynamics and mechanistic interpretability - Multimodal and multiview learning - Disentangled representations - Linear mode connectivity - Representation Similarity Analysis (RSA) - Multi-agent communication - Mechanistic interpretability - Representation universality Submission Tracks Full Paper (Archival) - Up to 9 pages (excluding references and appendix) - Anonymous submission - Published in dedicated workshop proceedings - Poster + spotlight opportunities Extended Abstract (Non-Archival) - Up to 4 pages - Anonymous submission - Poster + spotlight opportunities - Perfect for early-stage work, negative results, or opinion pieces Both tracks are eligible for best papers awards. Submit via OpenReview <https://openreview.net/group?id=UniReps.org/2026/Workshop> Deadline: 4th October 2026 (AoE) Join Our Program Committee We’re looking for dedicated reviewers to help maintain the high standard of our review process. Each reviewer is assigned no more than two submissions, and we’re planning to introduce Reviewer Awards to recognise outstanding contributions. Interested? Sign up here <https://docs.google.com/forms/d/e/1FAIpQLScGnnXH13fNw50Sm09oeXvTup-ZkpmSLq42...> to join the UniReps 2026 Program Committee. We look forward to your submissions and hope to see you at UniReps! Warm regards, The UniReps Organising Team