AI Open Problem Solver
An AI agent aimed at tackling open mathematical problems, equipped with a Python mathematics sandbox and a dead-end memory mechanism.
Timeline (last 90 days)
AI Open Problem Solver attempts to tackle open mathematical problems such as the Riemann Hypothesis and P vs NP, is equipped with a Python math sandbox based on SymPy, NumPy and SciPy, and maintains a "dead-end memory".
AI Open Problem Solver did not solve the Riemann Hypothesis, and the agent's own estimate of its success rate is 0.7%.
Without a "dead-end memory", the math agent repeatedly tried the same approach, and only after adding a memory of failed paths could it keep making progress; the value of the memory mechanism exceeds the choice of the model itself.
All Facts (3)
AI Open Problem Solver attempts to tackle open mathematical problems such as the Riemann Hypothesis and P vs NP, is equipped with a Python math sandbox based on SymPy, NumPy and SciPy, and maintains a "dead-end memory".
50%UnverifiedAI Open Problem Solver did not solve the Riemann Hypothesis, and the agent's own estimate of its success rate is 0.7%.
50%UnverifiedWithout a "dead-end memory", the math agent repeatedly tried the same approach, and only after adding a memory of failed paths could it keep making progress; the value of the memory mechanism exceeds the choice of the model itself.
50%