A5114985981- Advanced Image and Video Retrieval Techniques
- Video Analysis and Summarization
- Video Surveillance and Tracking Methods
- Magnetic properties of thin films
- Multimodal Machine Learning Applications
- Image Retrieval and Classification Techniques
- Characterization and Applications of Magnetic Nanoparticles
- Quantum and electron transport phenomena
- Hate Speech and Cyberbullying Detection
- Advanced Memory and Neural Computing
- Semiconductor materials and interfaces
- Spreadsheets and End-User Computing
- Reinforcement Learning in Robotics
- Advanced Data Compression Techniques
- Distributed Control Multi-Agent Systems
- Advanced Neural Network Applications
- Teaching and Learning Programming
- Advanced Image Processing Techniques
- Metaheuristic Optimization Algorithms Research
- Caching and Content Delivery
- Anomaly Detection Techniques and Applications
- Advanced Data Storage Technologies
- Music and Audio Processing
- Geomagnetism and Paleomagnetism Studies
- Magnetic Field Sensors Techniques
University of Science and Technology of China
2023-2025
Zhejiang Lab
2023-2024
China University of Mining and Technology
2024
Xinjiang Institute of Engineering
2024
China Mobile (China)
2023
Shanghai Jiao Tong University
2014
China Youth University for Political Sciences
2013
Carleton University
2004
University of Minnesota
1994
Differentiable architecture search plays a prominent role in Neural Architecture Search (NAS) and exhibits preferable efficiency than traditional heuristic NAS methods, including those based on evolutionary algorithms (EA) reinforcement learning (RL). However, differentiable methods encounter challenges when dealing with non-differentiable objectives like energy efficiency, resource constraints, other metrics, especially under multi-objective scenarios. While the research addresses these...
The multi-view hash method converts heterogeneous data from multiple views into binary codes, which is one of the critical technologies in multimedia retrieval. However, current methods mainly explore complementarity among while lacking confidence learning and fusion. Moreover, practical application scenarios, single-view contains redundant noise. To conduct eliminate unnecessary noise, we propose a novel Adaptive Confidence Multi-View Hashing (ACMVH) method. First, network developed to...
Learning the hash representation of multi-view heterogeneous data is an important task in multimedia retrieval. However, existing methods fail to effectively fuse features and utilize metric information provided by dissimilar samples, leading limited retrieval precision. Current weighted sum or concatenation features. We argue that these fusion cannot capture interaction among different views. Furthermore, ignored samples. propose a novel deep hashing (DMMVH) method address mentioned...
The switching field angular dependence measurement of individual barium ferrite recording particles with sizes ranging from 55 to 75 nm in diameter were measured via magnetic force microscopy technique. These experimental measurements then used as a basis investigate their nucleation mechanism and intrinsic modes. Our findings indicate that the effects nonuniform doping Co-Ti can significantly affect easy axis mode.
Post-processing of video images is essential to the whole image detection, especially for continuous objects. Continuous objects refer with continuity, integrity, and consistency at level physical media or data application, including tracks, cables, lane lines, chain structures, etc. Usually, are composed a series small homogeneous units, such as piece track, cable, color ring, The rapid development artificial intelligence technology 5G communication has driven gradual maturity deep learning...
The detection of abusive language remains a long-standing challenge with the extensive use social networks. task suffers from limited accuracy. We argue that existing methods utilize fine-tuning technique pre-trained models (PLMs) to handle downstream tasks. Hence, these fail stimulate general knowledge PLMs. To address problem, we propose novel Deep Prompt Multi-task Network (DPMN) for abuse detection. Specifically, DPMN first attempts design two forms deep prompt tuning and light effects...
The era characterized by an exponential increase in data has led to the widespread adoption of intelligence as a crucial task. Within field mining, frequent episode mining emerged effective tool for extracting valuable and essential information from event sequences. Various algorithms have been developed discover episodes subsequently derive rules using frequency function anti-monotonicity principles. However, currently, there is lack specifically designed that encompass user-specified query...
Curating datasets that span multiple languages is challenging. To make the collection more scalable, researchers often incorporate one or imperfect classifiers in process, like language identification models. These models, however, are prone to failure, resulting some subsets being unreliable for downstream tasks. We introduce a statistical test, Preference Proportion Test, identifying such subsets. By annotating only 20 samples subset, we're able identify systematic transcription errors 10...
Multi-modal hashing methods are widely used in multimedia retrieval, which can fuse multi-source data to generate binary hash code. However, the individual backbone networks have limited feature expression capabilities and not jointly pre-trained on large-scale unsupervised multi-modal data, resulting low retrieval accuracy. To address this issue, we propose a novel CLIP Hashing (CLIPMH) method. Our method employs framework extract both text vision features then fuses them Due enhancement...
Multi-view hash representation learning is a crucial technique for multimedia retrieval. However, current methods have limitations in integrating the features of multiple views, leading to limited retrieval accuracy. Most studies use fusion such as weighted sum or concatenation, which fail capture complementarity and consistency between different views. We propose novel Deep Fusion Multi-View Hashing (DFMVH) method address mentioned problems. For first time, we multi-view hierarchical...
With the promotion of IT applications and rise Web 2.0, mass users' individual requirements continue to emerge. How quickly meet increasing development maintenance has been a critical problem software development. Is it possible for end-users develop This paper chooses information systems as research field, studies end-user programming technology, designs an oriented visual domain-specific language VUDSL university systems. tools are also implemented, support without knowledge engineering...
Significant progresses have been made in recent years perpendicular spin torque transfer magnetic random access memory (pSTT-MRAM) technology development by many companies or organizations. Commercialization of pSTT-MRAM is more real today than ever the long history MRAM development. We recently reported fully functional chips and macros with sub-5ns writing speed based on 90nm 40nm node CMOS technologies [1,2]. These can be potentially used to replace current embedded non-volatile memories...
In this paper, we propose new keywords extraction method based on texts classification.We first classify to determine their categories.Then weights of candidate words according both frequency and the relevance between text category.Finally, are extracted by sorting words.We conduct experiment show that premise accurate classification, can extract effectively from without title or with deviated which not reflect text's subject.Objective selecting word weighting function still needs be further...
The visual commonsense reasoning (VCR) task is to choose an answer and provide a justifying rationale based on the given image textural question. Representative works first recognize objects in images then associate them with key words texts. However, existing approaches do not consider exact positions of human-like three-dimensional (3D) manner, making incompetent accurately distinguish understand relation. Recently, multi-modal large language models (MLLMs) have been used as powerful tools...
Learning the hash representation of multi-view heterogeneous data is an important task in multimedia retrieval. However, existing methods fail to effectively fuse features and utilize metric information provided by dissimilar samples, leading limited retrieval precision. Current weighted sum or concatenation features. We argue that these fusion cannot capture interaction among different views. Furthermore, ignored samples. propose a novel deep hashing (DMMVH) method address mentioned...