The field of software engineering is witnessing significant advancements with the integration of large language models (LLMs). Recent developments indicate a shift towards leveraging LLMs for automated testing, code generation, and program repair. Notably, researchers are exploring the use of LLMs for unit test generation, with a focus on improving test diversity and coverage. Additionally, LLMs are being applied to enhance the reliability of software systems, including industrial robotic systems and formal specifications. The use of LLMs for automated issue resolution is also gaining traction, with techniques such as adversarial iterative refinement and intent-guided semantic retrieval showing promise. Overall, the field is moving towards more effective and efficient software engineering practices, with LLMs playing a key role in driving innovation. Noteworthy papers include HPCAgentTester, which introduces a novel multi-agent LLM framework for automated unit test generation, and InfCode, which presents an adversarial multi-agent framework for automated repository-level issue resolution. InfCode-C++ is also notable, as it achieves a significant performance improvement on the C++ subset of MultiSWE-bench.
Advancements in Large Language Models for Software Engineering
Sources
Go-UT-Bench: A Fine-Tuning Dataset for LLM-Based Unit Test Generation in Go
HPCAgentTester: A Multi-Agent LLM Approach for Enhanced HPC Unit Test Generation
Automata-Based Steering of Large Language Models for Diverse Structured Generation
Beyond Accuracy: Behavioral Dynamics of Agentic Multi-Hunk Repair
WITNESS: A lightweight and practical approach to fine-grained predictive mutation testing
Actionable Warning Is Not Enough: Recommending Valid Actionable Warnings with Weak Supervision
SeedAIchemy: LLM-Driven Seed Corpus Generation for Fuzzing
Enhancing LLM Code Generation Capabilities through Test-Driven Development and Code Interpreter
SAINT: Service-level Integration Test Generation with Program Analysis and LLM-based Agents
FlakyGuard: Automatically Fixing Flaky Tests at Industry Scale
KTester: Leveraging Domain and Testing Knowledge for More Effective LLM-based Test Generation
Mutation Testing for Industrial Robotic Systems
Finetuning LLMs for Automatic Form Interaction on Web-Browser in Selenium Testing Framework
MutDafny: A Mutation-Based Approach to Assess Dafny Specifications
Quantum-Guided Test Case Minimization for LLM-Based Code Generation
Technique to Baseline QE Artefact Generation Aligned to Quality Metrics
Sequential testing problem: A follow-up review
TB or Not TB: Coverage-Driven Direct Preference Optimization for Verilog Stimulus Generation
InfCode: Adversarial Iterative Refinement of Tests and Patches for Reliable Software Issue Resolution
InfCode-C++: Intent-Guided Semantic Retrieval and AST-Structured Search for C++ Issue Resolution
An Agent-Based Framework for the Automatic Validation of Mathematical Optimization Models