Gregory Gay

Afonso Fontes, Gregory Gay

Context: Machine learning (ML) may enable effective automated test generation. Objective: We characterize emerging research, examining testing practices, researcher goals, ML techniques applied, evaluation, and challenges. Methods: We perform a systematic mapping on a sample of 124 publications. Results: ML generates input for system, GUI, unit, performance, and combinatorial testing or improves the performance of existing generation methods. ML is also used to generate test verdicts, property-based, and expected output oracles. Supervised learning - often based on neural networks - and reinforcement learning - often based on Q-learning - are common, and some publications also employ unsupervised or semi-supervised learning. (Semi-/Un-)Supervised approaches are evaluated using both traditional testing metrics and ML-related metrics (e.g., accuracy), while reinforcement learning is often evaluated using testing metrics tied to the reward function. Conclusion: Work-to-date shows great promise, but there are open challenges regarding training data, retraining, scalability, evaluation complexity, ML algorithms employed - and how they are applied - benchmarks, and replicability. Our findings can serve as a roadmap and inspiration for researchers in this field.

Jingxiong Liu, Ludvig Lemner, Linnea Wahlgren et al.

Much of the cost and effort required during the software testing process is invested in performing test maintenance - the addition, removal, or modification of test cases to keep the test suite in sync with the system-under-test or to otherwise improve its quality. Tool support could reduce the cost - and improve the quality - of test maintenance by automating aspects of the process or by providing guidance and support to developers. In this study, we explore the capabilities and applications of large language models (LLMs) - complex machine learning models adapted to textual analysis - to support test maintenance. We conducted a case study at Ericsson AB where we explore the triggers that indicate the need for test maintenance, the actions that LLMs can take, and the considerations that must be made when deploying LLMs in an industrial setting. We also propose and demonstrate a multi-agent architecture that can predict which tests require maintenance following a change to the source code. Collectively, these contributions advance our theoretical and practical understanding of how LLMs can be deployed to benefit industrial test maintenance processes.

Lirong Yi, Gregory Gay, Philipp Leitner

Large Language Models (LLMs) can generate code, but can they generate fast code? In this paper, we study this question using a dataset of 65 real-world tasks mined from open-source Java programs. We specifically select tasks where developers achieved significant speedups, and employ an automated pipeline to generate patches for these issues using two leading LLMs under four prompt variations. By rigorously benchmarking the results against the baseline and human-authored solutions, we demonstrate that LLM-generated code indeed improves performance over the baseline in most cases. However, patches proposed by human developers outperform LLM fixes by a statistically significant margin, indicating that LLMs often fall short of finding truly optimal solutions. We further find that LLM solutions are semantically identical or similar to the developer optimization idea in approximately two-thirds of cases, whereas they propose a more original idea in the remaining one-third. However, these original ideas only occasionally yield substantial performance gains.

Afonso Fontes, Gregory Gay, Francisco Gomes de Oliveira Neto et al.

Unit testing is a stage of testing where the smallest segment of code that can be tested in isolation from the rest of the system - often a class - is tested. Unit tests are typically written as executable code, often in a format provided by a unit testing framework such as pytest for Python. Creating unit tests is a time and effort-intensive process with many repetitive, manual elements. To illustrate how AI can support unit testing, this chapter introduces the concept of search-based unit test generation. This technique frames the selection of test input as an optimization problem - we seek a set of test cases that meet some measurable goal of a tester - and unleashes powerful metaheuristic search algorithms to identify the best possible test cases within a restricted timeframe. This chapter introduces two algorithms that can generate pytest-formatted unit tests, tuned towards coverage of source code statements. The chapter concludes by discussing more advanced concepts and gives pointers to further reading for how artificial intelligence can support developers and testers when unit testing software.

Hamid Ebadi, Mahshid Helali Moghadam, Markus Borg et al.

With the growing capabilities of autonomous vehicles, there is a higher demand for sophisticated and pragmatic quality assurance approaches for machine learning-enabled systems in the automotive AI context. The use of simulation-based prototyping platforms provides the possibility for early-stage testing, enabling inexpensive testing and the ability to capture critical corner-case test scenarios. Simulation-based testing properly complements conventional on-road testing. However, due to the large space of test input parameters in these systems, the efficient generation of effective test scenarios leading to the unveiling of failures is a challenge. This paper presents a study on testing pedestrian detection and emergency braking system of the Baidu Apollo autonomous driving platform within the SVL simulator. We propose an evolutionary automated test generation technique that generates failure-revealing scenarios for Apollo in the SVL environment. Our approach models the input space using a generic and flexible data structure and benefits a multi-criteria safety-based heuristic for the objective function targeted for optimization. This paper presents the results of our proposed test generation technique in the 2021 IEEE Autonomous Driving AI Test Challenge. In order to demonstrate the efficiency and effectiveness of our approach, we also report the results from a baseline random generation technique. Our evaluation shows that the proposed evolutionary test case generator is more effective at generating failure-revealing test cases and provides higher diversity between the generated failures than the random baseline.

Alireza Salahirad, Gregory Gay, Ehsan Mohammadi

Background: The field of software testing is growing and rapidly-evolving. Aims: Based on keywords assigned to publications, we seek to identify predominant research topics and understand how they are connected and have evolved. Method: We apply co-word analysis to map the topology of testing research as a network where author-assigned keywords are connected by edges indicating co-occurrence in publications. Keywords are clustered based on edge density and frequency of connection. We examine the most popular keywords, summarize clusters into high-level research topics, examine how topics connect, and examine how the field is changing. Results: Testing research can be divided into 16 high-level topics and 18 subtopics. Creation guidance, automated test generation, evolution and maintenance, and test oracles have particularly strong connections to other topics, highlighting their multidisciplinary nature. Emerging keywords relate to web and mobile apps, machine learning, energy consumption, automated program repair and test generation, while emerging connections have formed between web apps, test oracles, and machine learning with many topics. Random and requirements-based testing show potential decline. Conclusions: Our observations, advice, and map data offer a deeper understanding of the field and inspiration regarding challenges and connections to explore.

Afonso Fontes, Gregory Gay

Machine learning may enable the automated generation of test oracles. We have characterized emerging research in this area through a systematic literature review examining oracle types, researcher goals, the ML techniques applied, how the generation process was assessed, and the open research challenges in this emerging field. Based on a sample of 22 relevant studies, we observed that ML algorithms generated test verdict, metamorphic relation, and - most commonly - expected output oracles. Almost all studies employ a supervised or semi-supervised approach, trained on labeled system executions or code metadata - including neural networks, support vector machines, adaptive boosting, and decision trees. Oracles are evaluated using the mutation score, correct classifications, accuracy, and ROC. Work-to-date show great promise, but there are significant open challenges regarding the requirements imposed on training data, the complexity of modeled functions, the ML algorithms employed - and how they are applied - the benchmarks used by researchers, and replicability of the studies. We hope that our findings will serve as a roadmap and inspiration for researchers in this field.

Hussein Almulla, Gregory Gay

Search-based test generation is guided by feedback from one or more fitness functions - scoring functions that judge solution optimality. Choosing informative fitness functions is crucial to meeting the goals of a tester. Unfortunately, many goals - such as forcing the class-under-test to throw exceptions, increasing test suite diversity, and attaining Strong Mutation Coverage - do not have effective fitness function formulations. We propose that meeting such goals requires treating fitness function identification as a secondary optimization step. An adaptive algorithm that can vary the selection of fitness functions could adjust its selection throughout the generation process to maximize goal attainment, based on the current population of test suites. To test this hypothesis, we have implemented two reinforcement learning algorithms in the EvoSuite unit test generation framework, and used these algorithms to dynamically set the fitness functions used during generation for the three goals identified above. We have evaluated our framework, EvoSuiteFIT, on a set of Java case examples. EvoSuiteFIT techniques attain significant improvements for two of the three goals, and show limited improvements on the third when the number of generations of evolution is fixed. Additionally, for two of the three goals, EvoSuiteFIT detects faults missed by the other techniques. The ability to adjust fitness functions allows strategic choices that efficiently produce more effective test suites, and examining these choices offers insight into how to attain our testing goals. We find that adaptive fitness function selection is a powerful technique to apply when an effective fitness function does not already exist for achieving a testing goal.

Yang Ren, Gregory Gay, Christian Kästner et al.

Modern systems are built using development frameworks. These frameworks have a major impact on how the resulting system executes, how configurations are managed, how it is tested, and how and where it is deployed. Machine learning (ML) frameworks and the systems developed using them differ greatly from traditional frameworks. Naturally, the issues that manifest in such frameworks may differ as well---as may the behavior of developers addressing those issues. We are interested in characterizing the system-related issues---issues impacting performance, memory and resource usage, and other quality attributes---that emerge in ML frameworks, and how they differ from those in traditional frameworks. We have conducted a moderate-scale exploratory study analyzing real-world system-related issues from 10 popular machine learning frameworks. Our findings offer implications for the development of machine learning systems, including differences in the frequency of occurrence of certain issue types, observations regarding the impact of debate and time on issue correction, and differences in the specialization of developers. We hope that this exploratory study will enable developers to improve their expectations, plan for risk, and allocate resources accordingly when making use of the tools provided by these frameworks to develop ML-based systems.