From Defects to Demands: A Unified, Iterative, and Heuristically Guided LLM-Based Framework for Automated Software Repair and Requirement Realization
Alex, Liu, Vivian, Chi
TL;DR
Addressing the challenge of automated software repair and requirement realization, the paper proposes an LLM-driven iterative framework that combines test-driven development, formal verification, and heuristic search to refine a codebase from defects to demands. It formalizes code as a hypothesis space $\\mathcal{H}$ and a dynamic specification $\\varphi$, with a multi-channel feedback oracle producing a composite error $\\delta(C,\\varphi)$ defined by $\\delta(C,\\varphi) = \alpha_1 \\epsilon_{test} + \\alpha_2 \\epsilon_{struct} + \\alpha_3 \\epsilon_{verify} + \\alpha_4 \\epsilon_{logs}$. The framework provides convergence guarantees, including finite termination $\\lim_{t \to \infty} \\inf_C \\delta(C,\\varphi_t) = 0$ under reasonable assumptions, and employs a Gibbs-like update to monotone reduce $\\delta$. Empirically on SWE-bench, it achieves 67% acceptance versus 48.33% top prior, a 38.6% improvement, demonstrating autonomous AI-driven software engineering with robust context management and parallel verification; the work also analyzes ethical governance and outlines future directions.
Abstract
This manuscript signals a new era in the integration of artificial intelligence with software engineering, placing machines at the pinnacle of coding capability. We present a formalized, iterative methodology proving that AI can fully replace human programmers in all aspects of code creation and refinement. Our approach, combining large language models with formal verification, test-driven development, and incremental architectural guidance, achieves a 38.6% improvement over the current top performer's 48.33% accuracy on the SWE-bench benchmark. This surpasses previously assumed limits, signaling the end of human-exclusive coding and the rise of autonomous AI-driven software innovation. More than a technical advance, our work challenges centuries-old assumptions about human creativity. We provide robust evidence of AI superiority, demonstrating tangible gains in practical engineering contexts and laying the foundation for a future in which computational creativity outpaces human ingenuity.