Tao Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Tao Chen has a total of 38 co-authors. Here are the top co-authors by number of shared publications:
- Miqing Li
University of Birmingham, Birmingham, United Kingdom
8 shared publications - Jingzhi Gong
King's College London, London, United Kingdom
7 shared publications - Rami Bahsoon
University of Birmingham
7 shared publications - Xin Yao
Lingnan University
7 shared publications - Pengzhou Chen
School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, China
4 shared publications - Ke Li
Simon Fraser University
3 shared publications
Latest Publications
- Open Access
- 7 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 575-587
Less Is More: Tuning Configurable Systems with Imperfect Fidelity
- 7 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 575-587
Open AccessLess Is More: Tuning Configurable Systems with Imperfect Fidelity
- Open Access
- 28 August 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Evaluating Useful Surrogate Models for Configuration Tuning Beyond Accuracy: A Fitness Landscape Analysis Perspective
To efficiently tune configuration for better software system performance (e.g., latency) at the deployment and maintenance stage, many tuners have leveraged a surrogate model to expedite the process instead of solely relying on the profoundly expensive system measurement. As such, it is naturally believed that we need more accurate models. However, the fact of “ accuracy can lie ”—a somewhat surprising finding from prior work—has left us many unanswered questions regarding what role the surrogate model plays in configuration tuning. This paper provides the very first systematic exploration and discussion, together with a resolution proposal, to disclose the many faces of useful surrogate models for configuration tuning beyond accuracy, through the novel perspective of fitness landscape analysis. We present a theory as an alternative to accuracy for assessing the model usefulness in tuning, based on which we conduct an extensive empirical study involving up to \(27,000\) cases. Drawing on the above, we propose Model4Tune , an automated predictive tool that estimates which model-tuner pairs are the best for an unforeseen system without expensive tuner profiling. Our results suggest that Model4Tune , as one of the first of its kind, performs significantly better than random guessing in \(79\%-82\%\) of the cases, hence greatly mitigating the required efforts in engineering configuration for software systems. Our results not only shed light on the possible future research directions but also offer a practical resolution that can assist practitioners in evaluating the most useful model for configuration tuning.- 28 August 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Open AccessEvaluating Useful Surrogate Models for Configuration Tuning Beyond Accuracy: A Fitness Landscape Analysis Perspective
To efficiently tune configuration for better software system performance (e.g., latency) at the deployment and maintenance stage, many tuners have leveraged a surrogate model to expedite the process instead of solely relying on the profoundly expensive system measurement. As such, it is naturally believed that we need more accurate models. However, the fact of “ accuracy can lie ”—a somewhat surprising finding from prior work—has left us many unanswered questions regarding what role the surrogate model plays in configuration tuning. This paper provides the very first systematic exploration and discussion, together with a resolution proposal, to disclose the many faces of useful surrogate models for configuration tuning beyond accuracy, through the novel perspective of fitness landscape analysis. We present a theory as an alternative to accuracy for assessing the model usefulness in tuning, based on which we conduct an extensive empirical study involving up to \(27,000\) cases. Drawing on the above, we propose Model4Tune , an automated predictive tool that estimates which model-tuner pairs are the best for an unforeseen system without expensive tuner profiling. Our results suggest that Model4Tune , as one of the first of its kind, performs significantly better than random guessing in \(79\%-82\%\) of the cases, hence greatly mitigating the required efforts in engineering configuration for software systems. Our results not only shed light on the possible future research directions but also offer a practical resolution that can assist practitioners in evaluating the most useful model for configuration tuning. - Open Access
- 11 July 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (7) , 1-35
Causally Perturbed Fairness Testing
To mitigate unfair and unethical discrimination over sensitive features (e.g., gender, age, or race), fairness testing plays an integral role in engineering systems that leverage AI models to handle tabular data. A key challenge therein is how to effectively reveal fairness bugs under an intractable sample size using perturbation. Much current work has been focusing on designing the test sample generators, ignoring the valuable knowledge about data characteristics that can help guide the perturbation and hence limiting their full potential. In this article, we seek to bridge such a gap by proposing a generic framework of causally perturbed fairness testing, dubbed CausalFT . Through causal inference, the key idea of CausalFT is to extract the most directly and causally relevant non-sensitive feature to its sensitive counterpart, which can jointly influence the prediction of the label. Such a causal relationship is then seamlessly injected into the perturbation to guide a test sample generator. Unlike existing generator-level work, CausalFT serves as a higher-level framework that can be paired with diverse base generators. Extensive experiments on 1,296 cases confirm that CausalFT can considerably improve arbitrary base generators in revealing fairness bugs over 93% of the cases with acceptable extra run-time overhead. Compared with a state-of-the-art approach that ranks the non-sensitive features solely based on correlation, CausalFT performs significantly better on 64% cases while being much more efficient. Further, CausalFT can better improve bias resilience in nearly all cases.
- 11 July 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (7) , 1-35
Open AccessCausally Perturbed Fairness Testing
To mitigate unfair and unethical discrimination over sensitive features (e.g., gender, age, or race), fairness testing plays an integral role in engineering systems that leverage AI models to handle tabular data. A key challenge therein is how to effectively reveal fairness bugs under an intractable sample size using perturbation. Much current work has been focusing on designing the test sample generators, ignoring the valuable knowledge about data characteristics that can help guide the perturbation and hence limiting their full potential. In this article, we seek to bridge such a gap by proposing a generic framework of causally perturbed fairness testing, dubbed CausalFT . Through causal inference, the key idea of CausalFT is to extract the most directly and causally relevant non-sensitive feature to its sensitive counterpart, which can jointly influence the prediction of the label. Such a causal relationship is then seamlessly injected into the perturbation to guide a test sample generator. Unlike existing generator-level work, CausalFT serves as a higher-level framework that can be paired with diverse base generators. Extensive experiments on 1,296 cases confirm that CausalFT can considerably improve arbitrary base generators in revealing fairness bugs over 93% of the cases with acceptable extra run-time overhead. Compared with a state-of-the-art approach that ranks the non-sensitive features solely based on correlation, CausalFT performs significantly better on 64% cases while being much more efficient. Further, CausalFT can better improve bias resilience in nearly all cases.
- Top Cited PapersOpen Access
- 17 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (7) , 1-47
Learning Software Bug Reports: A Systematic Literature Review
The recent advancement of artificial intelligence, in particular Machine Learning (ML), has witnessed its significant growth in various software engineering research fields. Among them, bug report analysis is one of such examples as it aims to automatically understand, extract and correlate information from the reports with the help of ML approaches. Despite the importance of ML in automating and enhancing bug report analysis, a comprehensive review that systematically examines the state-of-the-art in this area is still lacking. In this article, we provide a systematic literature review on this promising research topic. Our review covers 1,825 papers, from which we extract 204 most relevant studies for detailed analysis. Based on the statistics and trends observed in these reviewed studies, we obtained seven key findings summarized as follows: (1) The extensive use of Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) and \( k \) -Nearest Neighbor ( \( k \) NN) for bug report analysis, noting the underutilization of more advanced models like BERT due to their complexity and computational demands. (2) Word2Vec and TF-IDF are the most common methods for feature representation, with a notable increase in deep learning-based methods in recent years. (3) Stop word removal is the most common preprocessing method, followed by tokenization and stemming. Structural methods surged post-2020. (4) Eclipse and Mozilla are the most frequently evaluated software projects, reflecting their prominence in the field. (5) Bug categorization is the most popular task, followed by bug localization, assignment, and severity/priority prediction, with a growing interest in bug report summarization driven by advancements in NLP. (6) Most studies focus on general bug types, but there is increasing attention on specific bugs such as non-functional and performance bugs. (7) Common evaluation metrics include F1-score, Recall, Precision, and Accuracy, but bug report related evaluation metrics have not received significant attention. The majority of studies prefer \( k \) -fold cross-validation for model evaluation. (8) Many studies lack robust statistical tests or effect size measurements. Finally, based on the key findings, we discover six promising future research directions, by which we hope, together with the findings, can offer useful insights to practitioners of this particular research direction.- 17 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
- Vol. 35 (7) , 1-47
Top Cited PapersOpen AccessLearning Software Bug Reports: A Systematic Literature Review
The recent advancement of artificial intelligence, in particular Machine Learning (ML), has witnessed its significant growth in various software engineering research fields. Among them, bug report analysis is one of such examples as it aims to automatically understand, extract and correlate information from the reports with the help of ML approaches. Despite the importance of ML in automating and enhancing bug report analysis, a comprehensive review that systematically examines the state-of-the-art in this area is still lacking. In this article, we provide a systematic literature review on this promising research topic. Our review covers 1,825 papers, from which we extract 204 most relevant studies for detailed analysis. Based on the statistics and trends observed in these reviewed studies, we obtained seven key findings summarized as follows: (1) The extensive use of Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) and \( k \) -Nearest Neighbor ( \( k \) NN) for bug report analysis, noting the underutilization of more advanced models like BERT due to their complexity and computational demands. (2) Word2Vec and TF-IDF are the most common methods for feature representation, with a notable increase in deep learning-based methods in recent years. (3) Stop word removal is the most common preprocessing method, followed by tokenization and stemming. Structural methods surged post-2020. (4) Eclipse and Mozilla are the most frequently evaluated software projects, reflecting their prominence in the field. (5) Bug categorization is the most popular task, followed by bug localization, assignment, and severity/priority prediction, with a growing interest in bug report summarization driven by advancements in NLP. (6) Most studies focus on general bug types, but there is increasing attention on specific bugs such as non-functional and performance bugs. (7) Common evaluation metrics include F1-score, Recall, Precision, and Accuracy, but bug report related evaluation metrics have not received significant attention. The majority of studies prefer \( k \) -fold cross-validation for model evaluation. (8) Many studies lack robust statistical tests or effect size measurements. Finally, based on the key findings, we discover six promising future research directions, by which we hope, together with the findings, can offer useful insights to practitioners of this particular research direction. - Open Access
- 27 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Human-Machine Co-Boosted Bug Report Identification with Mutualistic Neural Active Learning
Bug reports, encompassing a wide range of bug types, are crucial for maintaining software quality. However, the increasing complexity and volume of bug reports pose a significant challenge in sole manual identification and assignment to the appropriate teams for resolution, as dealing with all the reports is time-consuming and resource-intensive. In this paper, we introduce a cross-project framework, dubbed Mutualistic Neural Active Learning ( MNAL ), designed for automated and more effective identification of bug reports from GitHub repositories boosted by human-machine collaboration. MNAL utilizes a neural language model that learns and generalizes reports across different projects, coupled with active learning to form neural active learning. A distinctive feature of MNAL is the purposely crafted mutualistic relation between the machine learners (neural language model) and human labelers (developers) when enriching the knowledge learned. That is, the most informative human-labeled reports and their corresponding pseudo-labeled ones are used to update the model while those reports that need to be labeled by developers are more readable and identifiable, thereby enhancing the human-machine teaming therein. We evaluate MNAL using a dataset of \(1,275,881\) reports from over \(127,000\) software projects against the state-of-the-art approaches, baselines, and different variants. The results indicate that, remarkably, MNAL achieves up to 95.8% and 196.0% effort reduction in terms of readability and identifiability during human labeling, respectively, while resulting in a better performance (e.g., F1-score) in bug report identification. Additionally, our MNAL is model-agnostic since it is capable of improving the model performance with various underlying neural language models. To further verify the efficacy of our approach, we conducted a qualitative case study involving 10 human participants, who rate MNAL as being more effective while saving more time and monetary resources. The dataset and code are made publicly available at https://github.com/ideas-labo/MNAL .- 27 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Open AccessHuman-Machine Co-Boosted Bug Report Identification with Mutualistic Neural Active Learning
Bug reports, encompassing a wide range of bug types, are crucial for maintaining software quality. However, the increasing complexity and volume of bug reports pose a significant challenge in sole manual identification and assignment to the appropriate teams for resolution, as dealing with all the reports is time-consuming and resource-intensive. In this paper, we introduce a cross-project framework, dubbed Mutualistic Neural Active Learning ( MNAL ), designed for automated and more effective identification of bug reports from GitHub repositories boosted by human-machine collaboration. MNAL utilizes a neural language model that learns and generalizes reports across different projects, coupled with active learning to form neural active learning. A distinctive feature of MNAL is the purposely crafted mutualistic relation between the machine learners (neural language model) and human labelers (developers) when enriching the knowledge learned. That is, the most informative human-labeled reports and their corresponding pseudo-labeled ones are used to update the model while those reports that need to be labeled by developers are more readable and identifiable, thereby enhancing the human-machine teaming therein. We evaluate MNAL using a dataset of \(1,275,881\) reports from over \(127,000\) software projects against the state-of-the-art approaches, baselines, and different variants. The results indicate that, remarkably, MNAL achieves up to 95.8% and 196.0% effort reduction in terms of readability and identifiability during human labeling, respectively, while resulting in a better performance (e.g., F1-score) in bug report identification. Additionally, our MNAL is model-agnostic since it is capable of improving the model performance with various underlying neural language models. To further verify the efficacy of our approach, we conducted a qualitative case study involving 10 human participants, who rate MNAL as being more effective while saving more time and monetary resources. The dataset and code are made publicly available at https://github.com/ideas-labo/MNAL . - Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2722-2734
Light over Heavy: Automated Performance Requirements Quantification with Linguistic Inducement
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2722-2734
Open AccessLight over Heavy: Automated Performance Requirements Quantification with Linguistic Inducement
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 14-26
PromiseTune: Unveiling Causally Promising and Explainable Configuration Tuning
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 14-26
Open AccessPromiseTune: Unveiling Causally Promising and Explainable Configuration Tuning
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Dually Hierarchical Drift Adaptation for Online Configuration Performance Learning
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Open AccessDually Hierarchical Drift Adaptation for Online Configuration Performance Learning
- Open Access
- 26 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Revealing Domain-Spatiality Patterns for Configuration Tuning: Domain Knowledge Meets Fitness Landscapes
Configuration tuning for better performance is crucial in quality assurance. Yet, there has long been a mystery on tuners’ effectiveness, due to the black-box nature of configurable systems. Prior efforts predominantly adopt static domain analysis (e.g., static taint analysis), which often lacks generalizability, or dynamic data analysis (e.g., benchmarking performance analysis), limiting explainability. In this work, we embrace Fitness Landscape Analysis (FLA) as a bridge between domain knowledge and difficulty of the tuning. We propose Domland , a two-pronged methodology that synergizes the spatial information obtained from FLA and domain-driven analysis to systematically capture the hidden characteristics of configuration tuning cases, explaining how and why a tuner might succeed or fail. This helps to better interpret and contextualize the behavior of tuners and inform tuner design. To evaluate Domland , we conduct a case study of nine software systems and 93 workloads, from which we reveal several key findings: (1) configuration landscapes are inherently system-specific, with no single domain factor (e.g., system area, programming language, or resource intensity) consistently shaping their structure; (2) the core options (e.g., pic-struct of x264), which control the main functional flows, exert a stronger influence on landscape ruggedness (i.e. the difficulty of tuning) compared to resource options (e.g., cpu-independent of x264); (3) Workload effects on landscape structure are not uniformly tied to type or scale. Both contribute to landscape variations, but their impact is system-dependent.
- 26 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Software Engineering and Methodology
Open AccessRevealing Domain-Spatiality Patterns for Configuration Tuning: Domain Knowledge Meets Fitness Landscapes
Configuration tuning for better performance is crucial in quality assurance. Yet, there has long been a mystery on tuners’ effectiveness, due to the black-box nature of configurable systems. Prior efforts predominantly adopt static domain analysis (e.g., static taint analysis), which often lacks generalizability, or dynamic data analysis (e.g., benchmarking performance analysis), limiting explainability. In this work, we embrace Fitness Landscape Analysis (FLA) as a bridge between domain knowledge and difficulty of the tuning. We propose Domland , a two-pronged methodology that synergizes the spatial information obtained from FLA and domain-driven analysis to systematically capture the hidden characteristics of configuration tuning cases, explaining how and why a tuner might succeed or fail. This helps to better interpret and contextualize the behavior of tuners and inform tuner design. To evaluate Domland , we conduct a case study of nine software systems and 93 workloads, from which we reveal several key findings: (1) configuration landscapes are inherently system-specific, with no single domain factor (e.g., system area, programming language, or resource intensity) consistently shaping their structure; (2) the core options (e.g., pic-struct of x264), which control the main functional flows, exert a stronger influence on landscape ruggedness (i.e. the difficulty of tuning) compared to resource options (e.g., cpu-independent of x264); (3) Workload effects on landscape structure are not uniformly tied to type or scale. Both contribute to landscape variations, but their impact is system-dependent.
- 24 June 2025
- Published by Springer Nature
- p. 1-15
Multimodal Emotion Recognition by Fusing Video Semantics in Video Learning Scenarios
- 24 June 2025
- Published by Springer Nature
- p. 1-15
Multimodal Emotion Recognition by Fusing Video Semantics in Video Learning Scenarios