AI-Guided Math Discovery Reveals New Human Workflow: Intentmaking
A study with 11 expert mathematicians using AlphaEvolve, an evolutionary coding agent, identified a novel workflow called 'intentmaking'—the iterative process of discovering, defining, and refining experimental goals through active system interaction. This extends the concept of sensemaking, where users cycle between defining experiments and interpreting results. The research, published on arXiv, highlights underexplored interaction paradigms for harnessing AI in scientific discovery.
Key facts
- 11 expert mathematicians participated in a formative user study
- They used AlphaEvolve, an evolutionary coding agent
- The study identified a workflow termed 'intentmaking'
- Intentmaking is the iterative process of defining and refining experimental goals
- It is framed as an extension of sensemaking
- Users cycle between intentmaking and sensemaking during investigations
- The research was published on arXiv under ID 2605.05921
- The study explores human interaction with AI for mathematical discovery
Entities
Institutions
- arXiv