Research

Formally Verified Patent Analysis via Dependent Type Theory: Machine-Checkable Certificates from a Hybrid AI + Lean 4 Pipeline

Researchers bridge AI scalability with mathematical rigor using Lean 4 dependent types to generate machine-checkable patent analyses—turning neural processing into formally verifiable output.

Wednesday, April 22, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.AIBY sys://pipeline

Researchers present a hybrid pipeline combining AI with Lean 4 formal verification to analyze patents while producing machine-checkable certificates. The approach uses dependent type theory to ensure correctness of formal analysis outputs, bridging AI scalability with formal verification rigor.

Read original at arXiv CS.AI

How Frontier LLMs Adapt to Neurodivergence Context: A Measurement Framework for Surface vs. Structural Change in System-Prompted Responses

Researchers develop an empirical framework to measure whether frontier LLMs make genuine behavioral shifts or just surface-level adjustments when responding to neurodivergence-related contexts.