We invite submissions of three types:

  1. Position or opinions papers on the topic of workshop (<=4 pages),
  2. Ongoing and unpublished work as extended abstracts (<=4 pages) (under review NeurIPS submissions are allowed in this track),
  3. Novel research focuses on the various aspects of responsible design of modality-specific and multimodal systems (long papers <=9 pages).

Submission site: https://openreview.net/group?id=NeurIPS.cc/2024/Workshop/RBFM

Latex Style: Workshop Template

Note: NeurIPS checklist is not mendatory and would not be counted towards the revewing process. However, we would appreciate the having one.

Specifically, we invite the submissions in the area of multi-modal foundational models and resonsibilty but not limited to:

  • Multi-modal Foundational Models (Discriminative and Generative) focusing on various modality such as Image / Video / Speech / Language (e.g., MLLMs, T2I or T2A diffusion models, etc.)
  • Reliability and security challenges
  • Adversarial attacks on multi-modal models
  • Training time strategy to ensure responsible behaviour
  • Data centric perspective for reliable models
  • Ethics, risks, and fairness of multi-modal models
  • Various multi-modal training/inference/evaluation efficiency aspects (e.g., handingling large resolution images/videos, long context video understanding, long-horizon high-res video generation)
  • Other responsibility aspects such as memorization, hallucinations, copyright, etc.
  • (Special track) Creating community guidelines for responsible design of next generation of foundational models
  • (Special track) Position papers on the theme of the workshop

Important Dates

  • Submission Deadline: September 14, 2024, AoE (UTC-12) Sep 20 2024 11:59AM UTC-0
  • Author Notification: October 14, 2024, AoE (UTC-12) October 9, 2024, AoE (UTC-12)
  • Camera-ready due: November 9, 2024, AoE (UTC-12)

Double-blind reviewing: The reviewing process will be double blind. Authors are responsible for anonymizing submissions. In particular, author names, affiliations, and acknowledgements should not be included in your submission. You should avoid providing any identifying information (even in the supplementary material).

Dual-submission policy: Submissions to this workshop is non-archival. We welcome submissions of ongoing and unpublished work, including papers that are under review at any venue. Work that have been previously published or accepted for publication at any venue will not be considered. Policy on dual submissions applies for the entire duration of the reviewing process.