Mathematical talent, quantifier elimination, and Noriko’s theorem. One either needs inspiration to cut in the right theorem, or lots of time enumerate cut-free proofs. :D

Mirco Giacobbe presenting a nice work on finite automata, stochastic analysis and supermartingales at CAV. I also like his work on neural termination analysis and ranking functions from ESEC/FSE presented at VeriProP. x.com/arxiv_cslo/sta…

Noah Bertram from Cornell Computer Science presenting his work on verifying envy-freeness in a cake cutting protocol with SMT. Includes the first verified implementation of a 3-agent protocol from STOC '16 and a new 4-agent protocol. x.com/DO/status/1796…

Today’s CAV talks were all great. Enjoyed learning about certified model counting, the Lean proof assistant and of course, Calculemus! on formal methods for legal reasoning. For those joining the Mt. Royal hike, we’ll leave at 6:30am from the Omni. See ya! i-cav.org/2024/mont-roya…

Looking for a unified theory to describe PL syntax and semantics. Some have suggested CCGs or Montague grammars. Others say CMTT or HOAS. Considering Datalog and SemGuS. I also like Okhotin’s work because of its complexity and algebraic closure properties: arxiv.org/abs/2012.03538

Eran Malach now giving a mechanistic explanation of how transformers solve copying and retrieval in practice and some examples of length generalization in various algorithmic tasks, using RASP as a programming language. #SimonsLive

Fascinating talk by Jon Kleinberg describing a novel take on the language induction problem, with a rare example of a positive result in complexity theory: while language identification is impossible (due to Gold), language generation (without outputting the grammar) is possible.