Reproduce “Noisy Deductive Reasoning: How Humans Construct Math, and How Math Constructs Universes” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
David H. Wolpert, David Kinney
Mathematical reasoning is redefined as an inherently stochastic and error-prone process, proposing Noisy Deductive Reasoning (NDR) machines to formally model this aspect. This framework provides a Bayesian justification for common mathematical practices like valuing multiple proofs and suggests a Stochastic Mathematical Universe Hypothesis, where the universe's probabilistic nature reflects underlying mathematical stochasticity.
Key insights
- Mathematical reasoning, whether by humans or AI, is inherently stochastic and prone to errors due to the physical limitations of computational processes.
- Common mathematical practices, such as valuing multiple independent proofs or using abductive reasoning, are rational and justified within a probabilistic framework where truth is established with high probability.
- The 'Stochastic Mathematical Universe Hypothesis' suggests that the apparent probabilistic nature of quantum mechanics might be a reflection of the underlying stochasticity in mathematics itself.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.