ICLR 2025 MLMP Woorkshop — Deadline Extended

[Nikita Kazeev]

Dear multiscale scholars,

After listening to your feedback, we have decided to extend the deadline to Apr 27 14!

The good thing it's the Irreproducible Results Track, so the less experiments you run, the greater is the acceptance chance! Submit here.

Not all experiments go as planned—some go spectacularly off the rails. The Irreproducible Results Track is a lighthearted yet insightful space to celebrate those unpredictable, chaotic, and downright bizarre research findings that refuse to be replicated (no matter how hard you try).

🔬 What we're looking for:
- Experiments that produced exciting results… just once.
- Studies where minor tweaks led to completely different outcomes.
- Theoretical models that worked beautifully—until tested.
- Any research story that makes you laugh, cry, or question the very nature of science.

📌 Submissions should:
✅ Share the experimental setup and what should have happened.
✅ Highlight what actually happened (bonus points for confusion).
✅ Reflect on lessons learned—serious or otherwise.
✅ Embrace the joy (and occasional heartbreak) of scientific discovery.

This track is about learning from the unexpected, celebrating scientific curiosity, and having a good laugh at the beautifully chaotic process of research. Submission format is free: it can be a properly formatted article, a horrendously formatted MS Word article, a poster, a video, etc. To clarify, this will not result in a real "ICLR workshop paper" – but we will allocate one oral presentation slot. But only if there is a good enough submission. Runner-ups might get an extra poster slot.

Finally, uniquely in the history of ML conferences, if you are not satisfied with the review outcome, you are allowed to challenge the program chair to trial by combat.

💀Deadline: April 14 AoE
🔔Acceptance notification: April 16 AoE
⚔️Challenge to by combat deadline: April 20 Singapore

Best,
Nikita