Ze Shi Li

Yinan Wu, Ze Shi Li, Kathryn Thomasset Stolee et al.

Artificial Intelligence (AI) is reshaping how developers adopt software engineering practices, yet the multi-dimensional nature of developer-AI interaction remains under-explored. Prior studies have primarily examined dimensions observable from developer activities such as "Prompt crafting" and "Code Editing", overlooking how hidden intentions and emotional dimensions intertwine with concrete actions during AI-assisted programming. To understand this phenomenon, we conducted a mixed-methods study with 76 developers split into AI-assisted and non-AI groups. Each performed programming tasks (Python with API management or Java with SQL). Developers retrospectively labeled their self-reported intentions, tool-supported actions, and emotions from screen recordings, supplemented by surveys and interviews. Our user study resulted in a novel model named S-IASE with four dimensions to describe programming behavior: intention, action, supporting tool, and emotion for a given development state. Our analysis reveals aggregated and sequential behavioral patterns. For example, using AI assistants often makes developers more focused on actively creating code, evaluating, and verifying generated results. AI-assisted participants showed emotionally stable development flow, as opposed to non-AI-assisted participants who experienced more fluctuating emotions. Interviews revealed further nuance: some developers reported impostor-like feelings, expressing guilt or self-doubt about relying on AI. Our work bridges an important gap in understanding the complexities of developer-AI interaction in programming context.

Chenyu Wang, Zhou Yang, Yunbo Lyu et al.

Artificial Intelligence (AI) is now used across nearly every industry, making AI model quality essential for building reliable and trustworthy systems. Historically, correctness has been the main focus, but industry AI models must also satisfy many other important quality attributes. To understand how these attributes are perceived, the challenges they create, and the solutions used in practice, we identify nine key quality attributes and interview 15 AI practitioners from diverse backgrounds. The interviews show that practitioners prioritize attributes differently depending on context. For example, efficiency can matter more than correctness in real-time applications, while scalability and deployability are no longer seen as primary concerns. Data imbalance emerges as a major obstacle to maintaining model correctness and robustness, and practitioners commonly use mitigation strategies such as active learning. We validate our main findings with a survey of 50 practitioners, which shows that most of the findings are widely recognized. These results can help researchers focus on the attributes practitioners value most and avoid improving one attribute at the expense of others that are considered more critical.

Neil A. Ernst, Ahmed Musa Awon, Swapnil Hingmire et al.

Research software (also called scientific software) is essential for advancing scientific endeavours. Research software encapsulates complex algorithms and domain-specific knowledge and is a fundamental component of all science. A pervasive challenge in developing research software is technical debt, which can adversely affect reliability, maintainability, and scientific validity. Research software often relies on the initiative of the scientific community for maintenance, requiring diverse expertise in both scientific and software engineering domains. The extent and nature of technical debt in research software are little studied, in particular, what forms it takes, and what the science teams developing this software think about their technical debt. In this paper we describe our multi-method study examining technical debt in research software. We begin by examining instances of self-reported technical debt in research code, examining 28k code comments across nine research software projects. Then, building on our findings, we interview research software engineers and scientists about how this technical debt manifests itself in their experience, and what costs it has for research software and research outputs more generally. We identify nine types of self-admitted technical debt unique to research software, and four themes impacting this technical debt.

Ziliang Wang, Xiaohong Zhang, Kechi Zhang et al.

Individual objects, whether users or services, within a specific region often exhibit similar network states due to their shared origin from the same city or autonomous system (AS). Despite this regional network similarity, many existing techniques overlook its potential, resulting in subpar performance arising from challenges such as data sparsity and label imbalance. In this paper, we introduce the regional-based dual latent state learning network(R2SL), a novel deep learning framework designed to overcome the pitfalls of traditional individual object-based prediction techniques in Quality of Service (QoS) prediction. Unlike its predecessors, R2SL captures the nuances of regional network behavior by deriving two distinct regional network latent states: the city-network latent state and the AS-network latent state. These states are constructed utilizing aggregated data from common regions rather than individual object data. Furthermore, R2SL adopts an enhanced Huber loss function that adjusts its linear loss component, providing a remedy for prevalent label imbalance issues. To cap off the prediction process, a multi-scale perception network is leveraged to interpret the integrated feature map, a fusion of regional network latent features and other pertinent information, ultimately accomplishing the QoS prediction. Through rigorous testing on real-world QoS datasets, R2SL demonstrates superior performance compared to prevailing state-of-the-art methods. Our R2SL approach ushers in an innovative avenue for precise QoS predictions by fully harnessing the regional network similarities inherent in objects.

Ziliang Wang, Xiaohong Zhang, Ze Shi Li et al.

Accurate Quality of Service (QoS) prediction is essential for enhancing user satisfaction in web recommendation systems, yet existing prediction models often overlook feature noise, focusing predominantly on label noise. In this paper, we present the Probabilistic Deep Supervision Network (PDS-Net), a robust framework designed to effectively identify and mitigate feature noise, thereby improving QoS prediction accuracy. PDS-Net operates with a dual-branch architecture: the main branch utilizes a decoder network to learn a Gaussian-based prior distribution from known features, while the second branch derives a posterior distribution based on true labels. A key innovation of PDS-Net is its condition-based noise recognition loss function, which enables precise identification of noisy features in objects (users or services). Once noisy features are identified, PDS-Net refines the feature's prior distribution, aligning it with the posterior distribution, and propagates this adjusted distribution to intermediate layers, effectively reducing noise interference. Extensive experiments conducted on two real-world QoS datasets demonstrate that PDS-Net consistently outperforms existing models, achieving an average improvement of 8.91% in MAE on Dataset D1 and 8.32% on Dataset D2 compared to the ate-of-the-art. These results highlight PDS-Net's ability to accurately capture complex user-service relationships and handle feature noise, underscoring its robustness and versatility across diverse QoS prediction environments.

Ziliang Wang, Xiaohong Zhang, Ze Shi Li et al.

Accurate prediction of Quality of Service (QoS) metrics is fundamental for selecting and managing cloud based services. Traditional QoS models rely on manual feature engineering and yield only point estimates, offering no insight into the confidence of their predictions. In this paper, we propose QoSBERT, the first framework that reformulates QoS prediction as a semantic regression task based on pre trained language models. Unlike previous approaches relying on sparse numerical features, QoSBERT automatically encodes user service metadata into natural language descriptions, enabling deep semantic understanding. Furthermore, we integrate a Monte Carlo Dropout based uncertainty estimation module, allowing for trustworthy and risk-aware service quality prediction, which is crucial yet underexplored in existing QoS models. QoSBERT applies attentive pooling over contextualized embeddings and a lightweight multilayer perceptron regressor, fine tuned jointly to minimize absolute error. We further exploit the resulting uncertainty estimates to select high quality training samples, improving robustness in low resource settings. On standard QoS benchmark datasets, QoSBERT achieves an average reduction of 11.7% in MAE and 6.7% in RMSE for response time prediction, and 6.9% in MAE for throughput prediction compared to the strongest baselines, while providing well calibrated confidence intervals for robust and trustworthy service quality estimation. Our approach not only advances the accuracy of service quality prediction but also delivers reliable uncertainty quantification, paving the way for more trustworthy, data driven service selection and optimization.

Swapnil Hingmire, Ze Shi Li, Shiyu et al.

Interpretation of topics is crucial for their downstream applications. State-of-the-art evaluation measures of topic quality such as coherence and word intrusion do not measure how much a topic facilitates the exploration of a corpus. To design evaluation measures grounded on a task, and a population of users, we do user studies to understand how users interpret topics. We propose constructs of topic quality and ask users to assess them in the context of a topic and provide rationale behind evaluations. We use reflexive thematic analysis to identify themes of topic interpretations from rationales. Users interpret topics based on availability and representativeness heuristics rather than probability. We propose a theory of topic interpretation based on the anchoring-and-adjustment heuristic: users anchor on salient words and make semantic adjustments to arrive at an interpretation. Topic interpretation can be viewed as making a judgment under uncertainty by an ecologically rational user, and hence cognitive biases aware user models and evaluation frameworks are needed.

Colin Werner, Ze Shi Li, Derek Lowlind et al.

Non-functional requirements (NFR), which include performance, availability, and maintainability, are vitally important to overall software quality. However, research has shown NFRs are, in practice, poorly defined and difficult to verify. Continuous software engineering practices, which extend agile practices, emphasize fast paced, automated, and rapid release of software that poses additional challenges to handling NFRs. In this multi-case study we empirically investigated how three organizations, for which NFRs are paramount to their business survival, manage NFRs in their continuous practices. We describe four practices these companies use to manage NFRs, such as offloading NFRs to cloud providers or the use of metrics and continuous monitoring, both of which enable almost real-time feedback on managing the NFRs. However, managing NFRs comes at a cost as we also identified a number of challenges these organizations face while managing NFRs in their continuous software engineering practices. For example, the organizations in our study were able to realize an NFR by strategically and heavily investing in configuration management and infrastructure as code, in order to offload the responsibility of NFRs; however, this offloading implied potential loss of control. Our discussion and key research implications show the opportunities, trade-offs, and importance of the unique give-and-take relationship between continuous software engineering and NFRs. Research artifacts may be found at https://doi.org/10.5281/zenodo.3376342.

Omar Elazhary, Colin Werner, Ze Shi Li et al.

In 2006, Fowler and Foemmel defined ten core Continuous Integration (CI) practices that could increase the speed of software development feedback cycles and improve software quality. Since then, these practices have been widely adopted by industry and subsequent research has shown they improve software quality. However, there is poor understanding of how organizations implement these practices, of the benefits developers perceive they bring, and of the challenges developers and organizations experience in implementing them. In this paper, we discuss a multiple-case study of three small- to medium-sized companies using the recommended suite of ten CI practices. Using interviews and activity log mining, we learned that these practices are broadly implemented but how they are implemented varies depending on their perceived benefits, the context of the project, and the CI tools used by the organization. We also discovered that CI practices can create new constraints on the software process that hurt feedback cycle time. For researchers, we show that how CI is implemented varies, and thus studying CI (for example, using data mining) requires understanding these differences as important context for research studies. For practitioners, our findings reveal in-depth insights on the possible benefits and challenges from using the ten practices, and how project context matters.

Colin Werner, Ze Shi Li, Neil Ernst et al.

Building shared understanding of requirements is key to ensuring downstream software activities are efficient and effective. However, in continuous software engineering (CSE) some lack of shared understanding is an expected, and essential, part of a rapid feedback learning cycle. At the same time, there is a key trade-off with avoidable costs, such as rework, that come from accidental gaps in shared understanding. This trade-off is even more challenging for non-functional requirements (NFRs), which have significant implications for product success. Comprehending and managing NFRs is especially difficult in small, agile organizations. How such organizations manage shared understanding of NFRs in CSE is understudied. We conducted a case study of three small organizations scaling up CSE to further understand and identify factors that contribute to lack of shared understanding of NFRs, and its relationship to rework. Our in-depth analysis identified 41 NFR-related software tasks as rework due to a lack of shared understanding of NFRs. Of these 41 tasks 78% were due to avoidable (accidental) lack of shared understanding of NFRs. Using a mixed-methods approach we identify factors that contribute to lack of shared understanding of NFRs, such as the lack of domain knowledge, rapid pace of change, and cross-organizational communication problems. We also identify recommended strategies to mitigate lack of shared understanding through more effective management of requirements knowledge in such organizations. We conclude by discussing the complex relationship between shared understanding of requirements, rework and, CSE.

Ze Shi Li, Colin Werner, Neil Ernst et al.

The enactment of the General Data Protection Regulation (GDPR) in 2018 forced any organization that collects and/or processes EU-based personal data to comply with stringent privacy regulations. Software organizations have struggled to achieve GDPR compliance both before and after the GDPR deadline. While some studies have relied on surveys or interviews to find general implications of the GDPR, there is a lack of in-depth studies that investigate compliance practices and compliance challenges of software organizations. In particular, there is no information on small and medium enterprises (SMEs), which represent the majority of organizations in the EU, nor on organizations that practice continuous integration. Using design science methodology, we conducted an in-depth study over the span of 20 months regarding GDPR compliance practices and challenges in collaboration with a small, startup organization. We first identified our collaborator's business problems and then iteratively developed two artifacts to address those problems: a set of operationalized GDPR principles, and an automated GDPR tool that tests those GDPR-derived privacy requirements. This design science approach resulted in four implications for research and for practice. For example, our research reveals that GDPR regulations can be partially operationalized and tested through automated means, which improves compliance practices, but more research is needed to create more efficient and effective means to disseminate and manage GDPR knowledge among software developers.