LLM-Powered Test Case Generation for Detecting Bugs in Plausible Programs

2024-04-16

AI Summary

This work addresses the challenge of detecting subtle bugs in “plausible programs”—programs that pass all test cases yet contain latent defects. We propose TrickCatcher, a novel LLM-based framework featuring a three-stage collaborative detection paradigm: (1) semantic-preserving program mutation via LLMs; (2) specification-driven symbolic input generation guided by natural-language specifications; and (3) cross-version behavioral consistency checking to identify anomalous executions. Our key contribution is the first decoupling and joint verification of program mutation and specification-grounded input generation. Evaluated on TrickyBugs and EvalPlus, TrickCatcher achieves recall, precision, and F1 scores 1.80×, 2.65×, and 1.66× higher than state-of-the-art baselines, respectively. All code and data are publicly released.

Technology Category

Natural Language Processing: Code Generation / Program Synthesis from Natural LanguageMachine Learning: Large Multimodal Models (LMMs)Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved informationEconomics, Online Markets and Human Computation: Cost models of using LLMs in production systemsSemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Detecting tricky bugs in plausible programs, those that pass existing test suites yet still contain bugs, remains a significant challenge in software testing. To address this problem, we propose TrickCatcher, an LLM-powered approach to generating test cases for uncovering bugs in plausible programs. TrickCatcher operates in three stages: First, it uses an LLM to generate program variants based on the program under test (PUT) and its specification. Second, it employs an LLM to construct an input generator from the specification for producing test inputs. Finally, these inputs are executed on both the PUT and its program variants to detect inconsistencies in their outputs. We evaluate TrickCatcher on two datasets, TrickyBugs and EvalPlus, which include 366 human-written and 151 AI-generated plausible programs with tricky bugs. TrickCatcher achieves recall, precision, and F1 scores that are 1.80x, 2.65x, and 1.66x those of the state-of-the-art baselines, respectively. Code and data used are available at https://github.com/RinCloud/TrickCatcher.

Problem

• Detecting tricky bugs in plausible programs passing existing tests
• Generating test cases using LLMs to uncover hidden bugs
• Improving recall and precision in bug detection over baselines

Innovation

• LLM generates program variants for testing
• LLM constructs input generator from specifications
• Detects inconsistencies between PUT and variants