Apriori Knowledge in an Era of Computational Opacity: The Role of AI in Mathematical Discovery
Abstract: Can we acquire apriori knowledge of mathematical facts from the outputs of computer programs? People like Burge have argued (correctly in our opinion) that, for example, Appel and Haken acquired apriori knowledge of the Four Color Theorem from their computer program insofar as their program simply automated human forms of mathematical reasoning. However, unlike such programs, we argue that the opacity of modern LLMs and DNNs creates obstacles in obtaining apriori mathematical knowledge from them in similar ways. We claim though that if a proof-checker automating human forms of proof-checking is attached to such machines, then we can obtain apriori mathematical knowledge from them after all, even though the original machines are entirely opaque to us and the proofs they output may not, themselves, be human-surveyable.
- Every planar map is four colorable, volume 98. American Mathematical Soc., 1989.
- Science in the age of large language models. Nature Reviews Physics, 5(5):277–280, 2023.
- Florian J Boge. Two dimensions of opacity and the deep learning predicament. Minds and Machines, 32(1):43–75, 2022.
- Tyler Burge. Computer proof, apriori knowledge, and other minds: The sixth philosophical perspectives lecture. Philosophical perspectives, 12:1–37, 1998.
- Kathleen A Creel. Transparency in complex computational systems. Philosophy of Science, 87(4):568–589, 2020.
- The four-color theorem and mathematical proof. The Journal of Philosophy, 77(12):803–820, 1980.
- Eamon Duede. Instruments, agents, and artificial intelligence: novel epistemic categories of reliability. Synthese, 200(6):491, 2022.
- Eamon Duede. Deep learning opacity in scientific discovery. Philosophy of Science, 90(5):1089–1099, 2023.
- Advancing mathematics by guiding human intuition with ai. Nature, 600(7887):70–74, 2021.
- The philosophy of simulation: hot new issues or same old stew? Synthese, 169(3):593–613, 2009.
- Joshua A. Grochow. New applications of the polynomial method: the cap set conjecture and beyond. Bull. Amer. Math. Soc., 56(1):29–64, 2019.
- Paul Humphreys. Extending ourselves: Computational science, empiricism, and scientific method. Oxford University Press, 2004.
- Paul Humphreys. The philosophical novelty of computer simulation methods. Synthese, 169(3):615–626, 2009.
- Philip Kitcher. Mathematical change and scientific change. New directions in the philosophy of mathematics, pages 215–242, 1998.
- Zachary C Lipton. The mythos of model interpretability: In machine learning, the concept of interpretability is both important and slippery. Queue, 16(3):31–57, 2018.
- David Marr. Vision: A computational investigation into the human representation and processing of visual information. MIT press, 2010.
- Mathematical discoveries from program search with large language models. Nature, pages 1–3, 2023.
- Solving olympiad geometry without human demonstrations. Nature, 625(7995):476–482, 2024.
- Thomas Tymoczko. The four-color problem and its philosophical significance. The journal of philosophy, 76(2):57–83, 1979.
- Thomas Tymoczko. New directions in the philosophy of mathematics: An anthology. Princeton University Press, 1998.
- Timothy Williamson. How deep is the distinction between a priori and a posteriori knowledge? In Albert Casullo and Joshua C. Thurow, editors, The A Priori in Philosophy, pages 291–312. Oxford University Press, 2013.
- John Zerilli. Explaining machine learning decisions. Philosophy of Science, 89(1):1–19, 2022.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.