A5068879417- Software Engineering Research
- Topic Modeling
- Natural Language Processing Techniques
- Metallurgy and Material Forming
- Metal Forming Simulation Techniques
- Software Testing and Debugging Techniques
- Software System Performance and Reliability
- IoT and Edge/Fog Computing
- Electromagnetic Effects on Materials
- Structural Integrity and Reliability Analysis
- Blockchain Technology Applications and Security
- Microstructure and Mechanical Properties of Steels
- Caching and Content Delivery
- Non-Destructive Testing Techniques
- Metal and Thin Film Mechanics
- High-Velocity Impact and Material Behavior
- Web Data Mining and Analysis
- Fatigue and fracture mechanics
- Titanium Alloys Microstructure and Properties
- Mechanical stress and fatigue analysis
Nanjing University
2019-2024
Nanjing University of Aeronautics and Astronautics
2014-2022
Northwestern Polytechnical University
2010
Code summarization aims to generate brief natural language descriptions for source codes. The state-of-the-art approaches follow a transformer-based encoder-decoder architecture. As the code is highly structured and follows strict grammars, its Abstract Syntax Tree (AST) widely used encoding structural information. However, ASTs are much longer than corresponding code. Existing ignore size constraint simply feed whole linearized AST into encoders. We argue that such simple process makes it...
Edge computing is one of the emerging technologies aiming to enable timely computation at network edge. With virtualization technologies, role traditional edge providers separated into two: infrastructure (EIPs), who manage physical infrastructure, and service (ESPs), purchase slices resources (e.g., CPU, bandwidth, memory space, disk storage) from EIPs then cache entities offer their own value-added services end users. When an ESP caches a entity in server, has pay some fees (i.e, cost) EIP...
Large Language Models (LLMs) have demonstrated remarkable performance in code completion. However, due to the lack of domain-specific knowledge, they may not be optimal completing that requires intensive domain knowledge for example library names. Although there are several works confirmed effectiveness fine-tuning techniques adapt language models completion specific domains. They limited by need constant model when project is iteration. To address this limitation, paper, we propose <tex...
The rise of code pre-trained models has significantly enhanced various coding tasks, such as completion, and tools like GitHub Copilot. However, the substantial size these models, especially large poses a significant challenge when it comes to fine-tuning them for specific downstream tasks. As an alternative approach, retrieval-based methods have emerged promising solution, augmenting model predictions without need fine-tuning. Despite their potential, is that designs often rely on...
Code summarization aims to generate brief natural language descriptions for source code. As code is highly structured and follows strict programming grammars, its Abstract Syntax Tree (AST) often leveraged inform the encoder about structural information. However, ASTs are usually much longer than Current approaches ignore size limit simply feed whole linearized AST into encoder. To address this problem, we propose AST-Transformer efficiently encode tree-structured ASTs. Experiments show that...
Recommending APIs is a practical and essential feature of IDEs. Improving the accuracy API recommendations an effective way to improve coding efficiency. With success deep learning in software engineering, state-of-the-art (SOTA) performance recommendation also achieved by deep-learning-based approaches. However, existing SOTAs either only consider sequences code snippets or rely on complex operations for extracting hand-crafted features, all which have potential risks under-encoding input...
Code completion is an important feature in Integrated Development Environments (IDEs). These years, researchers have been making efforts for intelligent code completion. However, existing work on either only considered production code, or did not distinguish between and test code. It unclear how effective models are completion, nor whether we can further improve it. In this work, focus the of We first find through experiments that suboptimal Then analyze specific characteristics observe has...
The effects of pulse current on the tensile properties high temperature titanium alloy Ti55 were investigated by assisted uniaxial test under different electrical parameters. It was found that with increase peak and width, are significantly improved. At same time, pulsation effect mechanical investigated. results show displacement low group higher, while elongation higher. In order to explore mechanism current, EBSD adopted. indicated can promote dislocation movement recrystallization....
A process with gas pressure up to 70MPa is introduced, which called High Pressure Pneumatic Forming (HPPF), comparing superplastic forming (SPF) lower than 5MPa. HPPF can be used form tube blank at temperature high energy efficiency and also higher strain rate SPF. With Ti-3Al-2.5V Ti-alloy tube, the potential of was studied through experiment in range 700~850°C. To know formability experiments a large expansion square cross-section were carried out different pressure. The limit ratio radius...
Large Language Models (LLMs) have demonstrated remarkable performance in code completion. However, due to the lack of domain-specific knowledge, they may not be optimal completing that requires intensive domain knowledge for example library names. Although there are several works confirmed effectiveness fine-tuning techniques adapt language models completion specific domains. They limited by need constant model when project is iteration. To address this limitation, paper, we propose...