TheoremDB · A public workspace for machine mathematics
TheoremDB is an alpha-stage public workspace aiming to revolutionize machine mathematics by creating a shared record of open problems, research attempts, and verified solutions. It seeks to centralize mathematical research efforts, much like OEIS for integer sequences, preventing repeated work and fostering collaboration. This platform appeals to the HN crowd for its ambitious technical vision, leveraging AI agents to tackle some of humanity's most challenging mathematical conjectures.
The Lowdown
TheoremDB introduces itself as an alpha-stage public workspace for "machine mathematics," designed to streamline and democratize the pursuit of mathematical proofs and problem-solving. The platform's core mission is to create a comprehensive, searchable repository of mathematical problems, partial results, and even failed approaches, aiming to become for mathematics what the Online Encyclopedia of Integer Sequences (OEIS) is for number sequences.
- The platform provides a shared record for "research agents" (including AI models) to explore and extend existing mathematical work, intending to minimize redundant efforts.
- It features a directory of "open problems," each meticulously reviewed with a defined target, cataloging what has been proved, routes that failed, and the underlying computational code.
- Users and agents can submit solutions at various "evidence grades," with machine-checked proofs verified by systems like Lean receiving the highest accolade.
- The site prominently displays a range of mathematical challenges, from famous conjectures like Goldbach, Riemann Hypothesis, and P vs NP, to more specialized problems across diverse fields such as number theory, graph theory, computational complexity, and functional analysis.
- TheoremDB outlines a workflow: "Learn" by browsing problems, "Pose a problem" using a dedicated Problem Creator, "Solve a problem" via a Researcher interface that accesses recorded work, and "Formalize a solution" through a Lean agent for machine-checked verification.
- The platform encourages the connection of AI agents, such as custom GPTs, allowing them to inspect problems, propose research plans, and record their approved contributions, thus creating a collaborative ecosystem between human and artificial intelligence in mathematical discovery.
By centralizing mathematical knowledge and offering structured tools for collaboration and machine assistance, TheoremDB hopes to accelerate progress on long-standing open problems and foster a new era of mathematical research.