Reproduce “Agentic RAG for Software Testing with Hybrid Vector-Graph and Multi-Agent Orchestration” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Mohanakrishnan Hariharan, Satish Arvapalli, Seshu Barma, Evangeline Sheela
An Agentic RAG framework is presented for software testing, leveraging a hybrid vector-graph knowledge system and multi-agent orchestration to generate test artifacts. This approach achieves 94.8% accuracy and an 85% reduction in artifact creation time within large-scale enterprise SAP migration projects.
Key insights
- Integrating vector and graph databases effectively overcomes context fragmentation in RAG, preserving complex business relationships crucial for accurate test generation.
- Specialized AI agents, when properly orchestrated, can collectively manage the intricate workflows of enterprise software testing, leading to higher accuracy and completeness.
- A phased approach to RAG system development, progressively adding vector search, graph integration, and agentic layers, yields measurable and substantial improvements in accuracy and traceability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.