Proving Bounds for the Randomized MaxCut Approximation Algorithm in Lean4
59 points
4 days ago
| 3 comments
| abhamra.com
| HN
sevensor
13 hours ago
[-]
I’m interested by this recent development where people are using LLMs with Lean to prove things that haven’t been formalized yet. This seems like a really good fit for LLMs. They’re good at translating approximate language to formal language, they’re good at information retrieval, and any failures are mitigated by Lean itself. I haven’t tried it, but I imagine failure modes might be proving useless lemmas along the way, taking a needlessly roundabout approach, or inadvertently proving something different than what you set out to do.
reply
bo1024
7 hours ago
[-]
Nice! I'd imagine that defining a cut as simply a partition of the vertices (rather than a subset of edges) may simplify things slightly.
reply
vatsachak
6 hours ago
[-]
Nice intro challenge
reply