NeurIPS Topic Selection

Use this skill before committing to NeurIPS. The target is not "any good AI paper"; it is a paper whose contribution will matter to the NeurIPS reviewer community and survive the current official track rules.

Fit signals

Strong NeurIPS candidates usually have one of these cores:

• a general ML method, model, objective, optimization, inference, or learning principle;
• a theory result that changes understanding of ML behavior or limits;
• a high-quality empirical finding about models, data, evaluation, robustness, or scaling;
• a use-inspired result with a real scientific, social, health, robotics, sustainability, or creative-AI problem and a clear ML contribution;
• a dataset, benchmark, or evaluation contribution that belongs in the correct current NeurIPS track rather than being forced into main track;
• a rigorous negative result that changes community understanding.