A5010572058- Software Engineering Research
- Software Engineering Techniques and Practices
- Open Source Software Innovations
- Scientific Computing and Data Management
- Topic Modeling
- Software System Performance and Reliability
- Software Reliability and Analysis Research
- Service-Oriented Architecture and Web Services
- Business Process Modeling and Analysis
- Web Data Mining and Analysis
- Model-Driven Software Engineering Techniques
- Semantic Web and Ontologies
- VLSI and FPGA Design Techniques
- Interconnection Networks and Systems
- Data Quality and Management
- Innovative Human-Technology Interaction
- Natural Language Processing Techniques
- Optical Network Technologies
- Neural Networks and Reservoir Computing
- Photonic and Optical Devices
- Mobile Crowdsensing and Crowdsourcing
- Intelligent Tutoring Systems and Adaptive Learning
- Low-power high-performance VLSI design
- Anomaly Detection Techniques and Applications
- Software Testing and Debugging Techniques
University of Science and Technology Beijing
2016-2025
McGill University
2018-2024
Southwest Jiaotong University
2020-2024
Polytechnique Montréal
2024
Morgan State University
2023-2024
Nanchang University
2024
Yunnan Agricultural University
2023
Xidian University
2019-2020
Yunnan University
2019-2020
Baidu (China)
2020
In most safety-critical domains the need for traceability is prescribed by certifying bodies. Trace links are generally created among requirements, design, source code, test cases and other artifacts, however, creating such manually time consuming error prone. Automated solutions use information retrieval machine learning techniques to generate trace links, current fail understand semantics of software artifacts or integrate domain knowledge into tracing process therefore tend deliver...
In our era of rapid technological advancement, the research landscape for writing assistants has become increasingly fragmented across various communities. We seek to address this challenge by proposing a design space as structured way examine and explore multidimensional intelligent interactive assistants. Through large community collaboration, we five aspects assistants: task, user, technology, interaction, ecosystem. Within each aspect, define dimensions (i.e., fundamental components an...
Software and systems traceability is widely accepted as an essential element for supporting many software development tasks. Today's version control provide inbuilt features that allow developers to tag each commit with one or more issue ID, thereby providing the building blocks from which project-wide can be established between feature requests, bug fixes, commits, source code, specific developers. However, our analysis of six open projects showed on average only 60% commits were linked...
Most modern Issue Tracking Systems (ITSs) for open source software (OSS) projects allow users to add comments issues. Over time, these accumulate into discussion threads embedded with rich information about the project, which can potentially satisfy diverse needs of OSS stakeholders. However, discovering and retrieving relevant from is a challenging task, especially when discussions are lengthy number issues in ITSs vast. In this paper, we address challenge by identifying types presented...
Human trajectory prediction is essential and promising in many related applications. This challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also surrounding environment. Recent works based on long-short term memory (LSTM) models have brought tremendous improvements task prediction. However, most them focus spatial influence humans ignore temporal influence. In this paper, we propose a novel spatial-temporal attention (ST-Attention) model,...
Abstract Photonic computation has garnered huge attention due to its great potential accelerate artificial neural network tasks at much higher clock rate digital electronic alternatives. Especially, reconfigurable photonic processor consisting of Mach–Zehnder interferometer (MZI) mesh is promising for matrix multiplier. It desired implement high-radix MZI boost the capability. Conventionally, three cascaded meshes (two universal N × unitary and one diagonal mesh) are needed express weight...
Reflecting on the limitations of a study is crucial part research process. In software engineering studies, this reflection typically conveyed through discussions or threats to validity. current practice, such seldom provide sufficient insight understand rationale for decisions taken before and during study, their implications. We revisit practice discussing validity identify its weaknesses. propose refocus self-reflection discussion centered notion trade-offs . argue that documenting allows...
The industrial 4.0 era is the fourth revolution and characterized by network penetration; therefore, traditional manufacturing value creation will undergo revolutionary changes. Artificial intelligence drive next technology revolution, knowledge graphs comprise main foundation of this revolution. intellectualization information an important part industry 4.0, we can efficiently integrate multisource heterogeneous data realize through powerful semantic association graphs. Knowledge have been...
This paper proposes an AI-based diagnostic method using MRI images for rotator cuff injuries to assist in treatment by segmenting tear areas and assessing severity. A multi-model deep learning network based on Unet + FPN architecture was developed automatically segment injury determine grades. dataset of 376 patients with 5640 used training, additional 94 1410 reserved testing. To optimize segmentation, a tailored matching strategy applied, achieving Intersection over Union (IoU) 0.79 ± 0.01...
In current practice, the behavior of Machine-Learned Components (MLCs) is not sufficiently specified by predefined requirements. Instead, they "learn" existing patterns from available training data, and make predictions for unseen data when deployed. On surface, their ability to extract behave accordingly specifically useful hard-to-specify concepts in certain safety critical domains (e.g., definition a pedestrian detection component vehicle). However, lack requirements specifications on...
The documentation practice for machine-learned (ML) models often falls short of established practices traditional software, which impedes model accountability and inadvertently abets inappropriate or misuse models. Recently, cards, a proposal documentation, have attracted notable attention, but their impact on the actual is unclear. In this work, we systematically study in field investigate how to encourage more responsible accountable practice. Our analysis publicly available cards reveals...
Domain modelling transforms domain problem descriptions written in natural language (NL) into analyzable and concise models (class diagrams) during requirements analysis or the early stages of design software development. Since practice requires time addition to skills experience, several approaches have been proposed automate semi-automate construction from expressed NL. Despite existing work on model extraction, some significant challenges remain unaddressed: (i) extracted are not accurate...
Inspiration from design examples plays a crucial role in the creative process of user interface design. However, current tools and techniques that support inspiration usually only focus on example browsing with limited control or similarity-based retrieval, leading to undesirable outcomes such as drift fixation. To address these issues, we propose GANSpiration approach suggests for both targeted serendipitous inspiration, leveraging style-based Generative Adversarial Network. A quantitative...
Contributors to open source software (OSS) communities assume diverse roles take different responsibilities. One major limitation of the current OSS tools and platforms is that they provide a uniform user interface regardless activities performed by various types contributors. This paper serves as non-trivial first step towards resolving this challenge demonstrating methodology establishing knowledge understand how contributors' their dynamics, reflected in contributors perform, are...
Traceability between published scientific breakthroughs and their implementation is essential, especially in the case of open-source software which implements bleeding-edge science its code. However, aligning link GitHub repositories academic papers can prove difficult, current practice establishing maintaining such links remains unknown. This paper investigates role references contained these repositories. We conduct a large-scale study 20 thousand that make to papers. use mixed-methods...
Business applications encode various business processes within an organization. process specification languages such as BPEL (Business Process Execution Language) are commonly used to integrate services in order automate To remain competitive edge, managers frequently modify their processes. Determining the cost of modifying a is not trivial since changes have account for source code services. In this paper, we propose approach estimating change service oriented application. The applies...
Attempts to utilize information retrieval techniques fully automate the creation of traceability links have been hindered by terminology mismatches between source and target artifacts. Therefore, current trace algorithms tend produce imprecise incomplete results. In this paper we address mismatch proposing an expert system which integrates a knowledge base domain concepts their relationships, set logic rules for defining relationships artifacts based on these rules, process mapping into...
As new requirements are introduced and implemented in a software system, developers must identify the set of source code classes which need to be changed. Therefore, past effort has focused on predicting impacted by requirement. In this paper, we introduce evaluate type information based intuition that associated with historical changes specific class likely exhibit semantic similarity impact class. This Requirements Set (R2RS) family metrics captures between requirement existing previously...
Software project artifacts such as source code, requirements, and change logs represent a gold-mine of actionable information. As result, software analytic solutions have been developed to mine repositories answer questions "who is the expert?," "which classes are fault prone?," or even domain experts for these fault-prone classes?" Analytics often require training configuring in order maximize performance within context each project. A cold-start problem exists when function applied without...
In open-source software (OSS), the design of usability is often influenced by discussions among community members on platforms such as issue tracking systems (ITSs). However, digesting rich information embedded in can be a major challenge due to vast number and diversity comments. We propose evaluate ArguLens, conceptual framework automated technique leveraging an argumentation model support effective understanding consolidation opinions ITSs. Through content analysis, we anatomized highly...
Tea is one of the most popular drinks in world. The rapid and accurate recognition tea diseases great significance for taking targeted preventive measures. In this paper, an information entropy masked vision transformation (IEM-ViT) model was proposed diseases. weighting (IEW) method used to calculate IE each segment image, so that could learn maximum amount knowledge more quickly accurately. An asymmetric encoder–decoder architecture autoencoder (MAE), where encoder operated on only a...
Data scientists reportedly spend a significant amount of their time in daily routines on data wrangling, i.e. cleaning and extracting features. However, wrangling code is often repetitive error-prone to write. Moreover, it easy introduce subtle bugs when reusing adopting existing code, which results reduced model quality. To support with we present technique generate documentation for code. We use (1) program synthesis techniques automatically summarize transformations (2) test case...