CodaMOSA: Escaping Coverage Plateaus in Test Generation with Pre-trained Large Language Models
Search-based software testing (SBST) generates high-coverage test cases for programs under test with a combination of test case generation and mutation. SBST’s performance relies on there being a reasonable probability of generating test cases that exercise the core of the program under test. Given such test cases, SBST can then explore the space around them to exercise various parts of the program. This paper explores whether Large Language Models (LLMs) of code, such as OpenAI’s Codex, can be used to help SBST’s exploration. Our proposed algorithm, CodaMOSA, conducts SBST until its coverage improvements stall, then asks Codex to provide example test cases for under-covered functions. These examples help SBST redirect its search to more useful areas of the search space. On an evaluation over 486 benchmarks, CodaMOSA achieves statistically significantly higher coverage on many more benchmarks (173 and 279) than it reduces coverage on (10 and 4), compared to SBST and LLM-only baselines.
Wed 17 MayDisplayed time zone: Hobart change
15:45 - 17:15 | Test generationSEIP - Software Engineering in Practice / DEMO - Demonstrations / Technical Track / NIER - New Ideas and Emerging Results / Journal-First Papers at Meeting Room 102 Chair(s): Chunyang Chen Monash University | ||
15:45 7mTalk | SoapOperaTG: A Tool for System Knowledge Graph Based Soap Opera Test Generation DEMO - Demonstrations Yanqi Su Australian National University, Zheming Han , Zhenchang Xing CSIRO’s Data61; Australian National University, Xiwei (Sherry) Xu CSIRO’s Data61, Liming Zhu CSIRO’s Data61, Qinghua Lu CSIRO’s Data61 | ||
15:52 7mTalk | GUI Testing to the Power of Parallel Q-Learning DEMO - Demonstrations Marco Mobilio University of Milano Bicocca, Diego Clerissi University of Milano-Bicocca, Giovanni Denaro University of Milano-Bicocca, Italy, Leonardo Mariani University of Milano-Bicocca | ||
16:00 15mTalk | BADGE: Prioritizing UI Events with Hierarchical Multi-Armed Bandits for Automated UI Testing Technical Track Dezhi Ran Peking University, Hao Wang Peking University, China, Wenyu Wang University of Illinois Urbana-Champaign, Tao Xie Peking University | ||
16:15 15mTalk | Efficiency Matters: Speeding Up Automated Testing with GUI Rendering Inference Technical Track Sidong Feng Monash University, Mulong Xie Australian National University, Chunyang Chen Monash University Pre-print | ||
16:30 15mTalk | CodaMOSA: Escaping Coverage Plateaus in Test Generation with Pre-trained Large Language Models Technical Track Caroline Lemieux University of British Columbia, Jeevana Priya Inala Microsoft Research, Shuvendu Lahiri Microsoft Research, Siddhartha Sen Microsoft Research | ||
16:45 15mTalk | Simulation-Driven Automated End-to-End Test and Oracle Inference SEIP - Software Engineering in Practice Shreshth Tuli Meta Platforms Inc. and Imperial College, Kinga Bojarczuk Facebook, Natalija Gucevska Facebook, Mark Harman University College London, Xiaoyu Wang Meta Platforms Inc., Graham Wright Meta Platforms Inc. | ||
17:00 7mTalk | Reasoning-Based Software Testing NIER - New Ideas and Emerging Results Luca Giamattei Università di Napoli Federico II, Roberto Pietrantuono Università di Napoli Federico II, Stefano Russo Università di Napoli Federico II Pre-print | ||
17:07 7mTalk | Automated Generation and Evaluation of JMH Microbenchmark Suites From Unit Tests Journal-First Papers Mostafa Jangali Concordia University, Yiming Tang Concordia University, Niclas Alexandersson Chalmers University of Technology, Philipp Leitner Chalmers University of Technology, Sweden / University of Gothenburg, Sweden, Jinqiu Yang Concordia University, Weiyi Shang University of Waterloo | ||