We welcome two categories of submissions:
  • 4-page submissions of technical contributions, benchmarks, negative results and opinion papers.
  • 1-page tiny papers (inspired by the ICLR tiny paper initiative) showcasing early research ideas to foster discussion and collaboration.
Submissions should be on any topics related to the Intrinsically Motivated Open-ended Learning themes across disciplines (machine learning, cognitive sciences, philosophy, psychology, and more). Examples of relevant topics include:
  • Intrinsic motivation in machines and animals,
  • Autonomous open-ended learning,
  • Autonomous goal generation,
  • Self-supervised representation learning,
  • Hierarchical and goal-directed RL,
  • Curriculum learning,
  • Interplay between intrinsic and extrinsic motivations,
  • Skill compositionality and chunking,
  • Knowledge transfer,
  • Avoiding catastrophic forgetting,
  • Learning neural/probabilistic abstract representations of state, actions and skills,
  • Open-ended development in children,
  • Role of metacognitive representations and processes in open-ended learning,
  • Playful exploration,
  • Language and sociality in open-ended learning,
  • Mitigation of risks of open-ended learning,
  • Philosophical and ethical implications of IMOL.

Key Dates

Submission Deadline Sept 16, 2024
Acceptance Notification Oct 9, 2024
Workshop Date Dec 14 or 15, 2024 (TBD)

All deadlines are end of day AoE (Anywhere on Earth).

Submission Instructions

Submissions must be made via OpenReview. All submissions must be anonymized. Full submissions may be up to 4 pages long (excluding references and appendices) and use the NeurIPS 2024 LaTeX style template without the checklist. Tiny papers should be a single page (excluding references and appendices), using the same template. Appendices can be added to the main PDF. Each paper will receive two reviews, and all reviews will be double-anonymized.

Full submissions should report original research, develop novel environments or benchmarks, highlight meaningful negative results, or provide opinions and synthesis of previous works. Authors can upload concise versions of parallel submissions to other conferences such as the NeurIPS main conference or ICLR. We accept dual submission but discourage submitting to multiple NeurIPS workshops.

All accepted papers will be available on the workshop website, but no formal workshop proceedings will be published.

For any questions, email us at imol.workshop@gmail.com