Reproduce “ResTest: Targeted Coverage of Residual Not-Covered Code Using Large Language Models for Web GUI Testing” 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
Chenxu Liu, Zhiyu Gu, Guoquan Wu, Ying Zhang, Wei Yang, Jun Wei, Tao Xie
ResTest augments web GUI testers by using coverage reports and a state-transition graph to propose workflows for an LLM-driven agent. It first records exploration, then targets under-covered files with tasks and related GUI states. Across ten applications, it improves coverage over three testers; against GPTWeb, coverage is 46.66% versus 52.75%, with a 17.52% relative gain across mixed coverage metrics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.