A5100413886- Software Engineering Research
- Video Analysis and Summarization
- Topic Modeling
- Advanced Vision and Imaging
- Human Pose and Action Recognition
- Brain Tumor Detection and Classification
- Parallel Computing and Optimization Techniques
- Advanced Image Processing Techniques
- Natural Language Processing Techniques
- Advanced Neural Network Applications
- Environmental Education and Sustainability
- Image Enhancement Techniques
- Software Testing and Debugging Techniques
- Advanced Computing and Algorithms
- AI in cancer detection
- Human Motion and Animation
- Rough Sets and Fuzzy Logic
- Advanced Battery Technologies Research
- Environmental Sustainability in Business
- Artificial Intelligence in Games
- Advancements in Battery Materials
- Image and Object Detection Techniques
- Software Reliability and Analysis Research
- Innovative Human-Technology Interaction
- Economic and Environmental Valuation
Beijing Foreign Studies University
2023
Huawei Technologies (Sweden)
2022
Huawei Technologies (United Kingdom)
2022
Beijing Institute of Technology
2017-2020
University of Waterloo
2020
Nanjing University of Posts and Telecommunications
2019
University of Science and Technology Beijing
2019
University of California, Davis
2017
Interface (United States)
2017
Automated generate-and-validate (GV) program repair techniques (APR) typically rely on hard-coded rules, thus only fixing bugs following specific fix patterns. These rules require a significant amount of manual effort to discover and it is hard adapt these different programming languages.
Layered lithium-rich cathode materials have been considered as competitive candidates for advanced lithium-ion batteries because they are environmentally benign, high capacity (more than 250 mAh·g–1), and low cost. However, still suffer from poor rate capability modest cycling performance. To address these issues, we proposed constructed a spinel-structure skin ferric oxide islands on the surface of layered through facile wet chemical method. During modification, Li ions in area pristine...
Abstract Traditional opto-ultrasonic grating instruments exhibit several limitations. Firstly, piezoelectric ceramics are prone to electrochemical corrosion, which shortens the instrument’s lifespan. Secondly, sound velocity in liquids is highly sensitive temperature variations, leading measurement discrepancies under different thermal conditions. Additionally, large size not only complicates its operation and handling but also indirectly affects accuracy. These issues collectively result...
Automatically generating compilable programs with (or without) natural language descriptions has always been a touchstone problem for computational linguistics and automated software engineering. Existing deep-learning approaches model code generation as text generation, either constrained by grammar structures in decoder, or driven pre-trained models on large-scale corpus (e.g., CodeGPT, PLBART, CodeT5). However, few of them account compilability the generated programs. To improve programs,...
Brain tumors have high morbidity and may lead to highly lethal cancer. In clinics, accurate segmentation of is the means for diagnosis determination subsequent treatment options. Due irregularity blurring tumor boundaries, accurately segmenting lesions has received extensive attention in medical image analysis. view this situation, paper proposed a brain method based on generative adversarial networks (GANs). The GAN architecture consists densely connected three-dimensional (3D) U-Net used...
This paper presents an effective approach for recognising human actions from depth video sequences by employing motion maps (DMMs) and convolutional neural networks (CNNs). Depth are projected onto three orthogonal planes, frame differences under each view (front/side/top) then accumulated through entire sequence generating a DMM. We build model architecture of multi-view network (MV-CNN) containing multiple to deal with DMMs (DMMf, DMMs, DMMt). The output full-connected layer is integrated...
Generating videos from text has proven to be a significant challenge for existing generative models. We tackle this problem by training conditional model extract both static and dynamic information text. This is manifested in hybrid framework, employing Variational Autoencoder (VAE) Generative Adversarial Network (GAN). The features, called "gist," are used sketch text-conditioned background color object layout structure. Dynamic features considered transforming input into an image filter....
Automatically generating compilable programs with (or without) natural language descriptions has always been a touchstone problem for computational linguistics and automated software engineering. Existing deep-learning approaches model code generation as text generation, either constrained by grammar structures in decoder, or driven pre-trained models on large-scale corpus (e.g., CodeGPT, PLBART, CodeT5). However, few of them account compilability the generated programs. To improve programs,...
Argument generation is an important but challenging task in computational argumentation.Existing studies have mainly focused on generating individual short arguments, while research long and coherent argumentative essays still under-explored.In this paper, we propose a new task, Argumentative Essay Generation (AEG).Given writing prompt, the goal of AEG to automatically generate essay with strong persuasiveness.We construct large-scale dataset, ArgEssay, for establish model based dual-decoder...
Arbitrary style transfer has drawn widespread attention from academic communities for its extensive application in reality. However, existing methods fail to make a proper trade-off between flexibility and capability. All of these directly synthesize textures on the whole content image indiscriminately, leading local distortions when there are obvious visual distinctions different regions. In this paper, we propose novel Saliency-Aware Instance Normalization (SAIN) module considering salient...
We propose a Leaked Motion Video Predictor (LMVP) to predict future frames by capturing the spatial and temporal dependencies from given inputs. The motion is modeled newly proposed component, guider, which plays role of both learner teacher. Specifically, it {\em learns} features real data guides} generator frames. consistency in video an adaptive filtering network. To further ensure spatio-temporal prediction, discriminator also adopted distinguish generated Further, leaks information...