Naive language models spark more creative scientific ideas

The Value of Not Knowing: How Fresh Eyes through Naive LLMs Fuel Creativity

Artificial IntelligenceComputation and LanguageMachine Learning

Summary

Sometimes, having less knowledge about a topic helps people see problems in new ways. The authors show that language models that are made 'naive'—deliberately lacking information about a subject—can come up with more original and useful ideas than usual models. These unique ideas can then be polished by more knowledgeable models. This approach suggests that knowing less can sometimes help generate truly outside-the-box solutions.

What this means in practice

  • For ai research teams: Generate novel research hypotheses by combining naive models' fresh ideas with mentor models' refinement.
  • For product innovation teams: Use naive models to seed unconventional product concepts that established AI tools might overlook.

Authors

David B. D'Ambrosio, Alon Albalak, Joel Lehman, Kenneth O. Stanley

Abstract

Many major scientific advances come from approaching a problem with a new or novel perspective. While large language models (LLMs) with internet-scale knowledge seem poised to take such perspectives, their consistent generation remains elusive. Instead, models are drawn to dominant views and preexisting, in-field bias. To explore the potential of fresh eyes to increase creativity, this paper proposes naive LLMs, models that are made intentionally ignorant of a specific topic. Through deprivation in training or intentional forgetting, naive models are unburdened by existing dogma, allowing them to be free to make new associations and conjectures. These naive models can then be paired with a stronger mentor model that can refine their unique perspectives into concrete proposals. For the 1B and 7B models tested in this paper, naive models consistently produce seed ideas that are more novel and useful than baseline counterparts, and mentor refinement improves the ideas even further. Thus, contrary to the prevailing tradition of scaling up training data, we demonstrate a case where less information is actually more powerful, in particular when the aim is to discover genuinely outside-the-box solutions.