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 improving code generation, automated bug fixing, and enhancing software development processes. Notably, LLMs are being used to generate high-quality code, detect bugs, and even repair them, showcasing their potential in reducing manual effort and improving software reliability. Furthermore, research is focusing on combining LLMs with other techniques such as retrieval-augmented generation and graph-based methods to enhance their capabilities. The use of LLMs in software engineering is expected to continue growing, with potential applications in areas like automated testing, code review, and project management. Some noteworthy papers include LLM-Based Program Generation for Triggering Numerical Inconsistencies Across Compilers, which presents a framework for generating programs to detect numerical inconsistencies, and RepoDebug, a repository-level multi-task and multi-language debugging evaluation of large language models. Another significant contribution is the development of TreeGPT, a novel hybrid architecture for abstract syntax tree processing, which has shown promising results in neural program synthesis tasks.
Advancements in Software Engineering with Large Language Models
Sources
LLM-Based Program Generation for Triggering Numerical Inconsistencies Across Compilers
JS-TOD: Detecting Order-Dependent Flaky Tests in Jest
Formalizing Linear Motion G-code for Invariant Checking and Differential Testing of Fabrication Tools
Compiler Bugs Detection in Logic Synthesis Tools via Linear Upper Confidence Bound
Generative Goal Modeling
Error Notebook-Guided, Training-Free Part Retrieval in 3D CAD Assemblies via Vision-Language Models
Aligning Requirement for Large Language Model's Code Generation
Automated Generation of Issue-Reproducing Tests by Combining LLMs and Search-Based Testing
Benchmarking and Studying the LLM-based Code Review
Automated Repair of C Programs Using Large Language Models
Scalable Thread-Safety Analysis of Java Classes with CodeQL
ReCode: Improving LLM-based Code Repair with Fine-Grained Retrieval-Augmented Generation
Are We SOLID Yet? An Empirical Study on Prompting LLMs to Detect Design Principle Violations
Analyzing Variations in Dependency Distributions Due to Code Smell Interactions
RepoDebug: Repository-Level Multi-Task and Multi-Language Debugging Evaluation of Large Language Models
Comparative Evaluation of Large Language Models for Test-Skeleton Generation
Code Review Without Borders: Evaluating Synthetic vs. Real Data for Review Recommendation
A Large-Scale Study of Floating-Point Usage in Statically Typed Languages
AI-Assisted Modeling: DSL-Driven AI Interactions
Multi-IaC-Eval: Benchmarking Cloud Infrastructure as Code Across Multiple Formats
Combining TSL and LLM to Automate REST API Testing: A Comparative Study
TreeGPT: A Novel Hybrid Architecture for Abstract Syntax Tree Processing with Global Parent-Child Aggregation
Natural Language-Programming Language Software Traceability Link Recovery Needs More than Textual Similarity
Automating API Documentation with LLMs: A BERTopic Approach
GRACE: Graph-Guided Repository-Aware Code Completion through Hierarchical Code Fusion
Proof2Silicon: Prompt Repair for Verified Code and Hardware Generation via Reinforcement Learning
Analyzing the Instability of Large Language Models in Automated Bug Injection and Correction
ChatGPT for Code Refactoring: Analyzing Topics, Interaction, and Effective Prompts
Design and Implementation of Code Completion System Based on LLM and CodeBERT Hybrid Subsystem
AutoVeriFix: Automatically Correcting Errors and Enhancing Functional Correctness in LLM-Generated Verilog Code
SWE-Mirror: Scaling Issue-Resolving Datasets by Mirroring Issues Across Repositories
Handling Open-Vocabulary Constructs in Formalizing Specifications: Retrieval-Augmented Parsing with Expert Knowledge