The paper introduces Lean4Agent, a framework utilizing Lean4 for formal modeling and verification of agent behavior, particularly for LLM-driven multi-step workflows. It addresses the absence of formal methods in current agent systems by allowing semantic consistency checks and enabling the localization of execution-time failures.
This research introduces a novel learning-to-refine framework to address the prohibitive computational cost of Large Language Models (LLMs) in formal theorem proving. By exploiting compiler outputs that compress diverse proof attempts into structured failure modes, the method enables efficient proof exploration and local error correction, significantly amplifying the reasoning capabilities of base provers.
This paper introduces a formally verified framework for patent analysis, employing a hybrid AI and Lean 4 pipeline. It marks the first known application of interactive theorem proving to intellectual property analysis, addressing the shortcomings of manual or probabilistic methods.
This is a v0.0 outline of the paper 'Rei as a Formal-Verification Compilation Pass for AI-Generated Mathematics', detailing Rei's proposal as a formal-verification compilation pass for AI-generated hypotheses. Current validation is at a scaffold-level smoke-run stage, with a full demonstration pending for v0.1 promotion.
This paper presents an algebraic semantics framework for governed execution, built on interaction trees and parameterized coinduction, with a mechanized implementation in Rocq. It introduces a GovernanceAlgebra that induces a symmetric monoidal category and an algebraic effect system ensuring governance-preserving handlers and capability-indexed composition.
This paper proposes BODHI, a domain knowledge prompting method for OS kernel specification inference, aiming to overcome current LLM limitations. It augments the standard few-shot prompt with a structured C-to-Python translation guide, improving automation and specification precision.
This research introduces Verifiable Transformers, a framework for converting task-localized Transformer circuits into bounded, solver-checkable claims. It enables verification of properties like functional equivalence and robustness through direct or surrogate-mediated SMT encoding.
NeuroNL2LTL is a neurosymbolic architecture that unifies learned translation with formal verification to translate natural language into Linear Temporal Logic. It employs verifier-in-the-loop training, where verification outcomes serve as reward signals for reinforcement learning, optimizing for formal correctness.