This paper investigates certified runtime monitoring of past-time signal temporal logic (ptSTL) from visual observations under partial observability. It proposes a reusable monitor that infers safety-relevant quantities from images and provides finite-sample guarantees, leveraging semantic latent representations to certify formulas without per-formula retraining.
Shrijith Venkatramana introduces git-lrc, an AI code reviewer, and discusses the growing complexity of authorization systems, especially with AI-generated code. He proposes using mathematical enforcement, like Lean 4, to ensure the safety of AI-generated code by making certain invalid states mechanically impossible.
This paper introduces ViTaX, a formal XAI framework designed to generate targeted semifactual explanations with mathematical guarantees. It addresses the shortcomings of existing XAI methods in providing trustworthy explanations for deep neural networks in safety-critical domains like autonomous driving and medical diagnosis.
This research presents a machine-checked formalization of AI workflow architectures with effect-transparent governance, demonstrating that governance can be imposed without losing computational expressivity. It defines a governance operator G for mediating effectful directives like memory access and LLM queries, proving seven key properties including governed Turing completeness and a decidability boundary.
The paper introduces a tool-mediated LLM architecture for autonomous cyber defense, designed to provide formal guarantees for high-stakes decision-making under adversarial pressure. It certifies controllability, observability, and Input-to-State Stability (ISS) robustness through a machine-checked Lyapunov function, demonstrating its effectiveness on real enterprise attack graphs.
This research introduces IC3-Evolve, a novel method for hardware model checking. It leverages proof- and witness-gated offline LLM-driven heuristic evolution to enhance the efficiency of the IC3 algorithm.
The title suggests an exploration of duality in reasoning about program execution, indicating a deep analysis of alternative approaches. It likely delves into formal and logical methods for understanding how programs operate.