New HyperFL framework enhances software fault localization with query-adaptive learning

Researchers have developed HyperFL, a novel framework designed to improve software fault localization by adapting representation learning to specific query characteristics. Unlike previous methods that use a fixed query representation, HyperFL utilizes a lightweight hypernetwork to generate query-specific LoRA parameters. This allows for dynamic adaptation of the query encoder while keeping the code encoder static and reusable. Experiments show HyperFL achieves significant performance gains, including up to a 16.7% relative improvement in Hit@1 over the state-of-the-art SweRank method. AI

IMPACT This research could lead to more efficient automated debugging and program repair by improving the accuracy of identifying code locations responsible for issues.

RANK_REASON The cluster contains a research paper detailing a new framework for software fault localization. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• LoRA
• ScienceCast
• SweRank

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →