A5100697598- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Machine Learning and Data Classification
- Generative Adversarial Networks and Image Synthesis
- Text and Document Classification Technologies
- Image Retrieval and Classification Techniques
- Face recognition and analysis
- Multimodal Machine Learning Applications
- Machine Learning and ELM
- Visual Attention and Saliency Detection
- Video Surveillance and Tracking Methods
- Advanced Image Processing Techniques
- Wastewater Treatment and Nitrogen Removal
- Molecular Sensors and Ion Detection
- Complex Network Analysis Techniques
- Image Processing and 3D Reconstruction
- Neural Networks and Applications
- Phosphorus and nutrient management
- Energy Efficient Wireless Sensor Networks
- Emotion and Mood Recognition
- Sulfur Compounds in Biology
- Traumatic Brain Injury Research
- Advanced Computing and Algorithms
- Impact of Light on Environment and Health
Capital Medical University
2024-2025
State Ethnic Affairs Commission
2023-2025
Huzhou University
2025
Yanshan University
2012-2024
South China University of Technology
2023-2024
Hohai University
2024
Beijing University of Technology
2024
Hefei University of Technology
2024
South Central Minzu University
2023
Qilu Hospital of Shandong University
2023
Knowledge distillation (KD) shows a bright promise as powerful regularization strategy to boost generalization ability by leveraging learned sample-level soft targets. Yet, employing complex pre-trained teacher network or an ensemble of peer students in existing KD is both timeconsuming and computationally costly. Various self methods have been proposed achieve higher efficiency. However, they either require extra architecture modification are difficult parallelize. To cope with these...
Accurately diagnosing patients with the vegetative state (VS) and minimally conscious (MCS) reached a misdiagnosis of approximately 40%.
We present the Recognize Anything Model (RAM): a strong foundation model for image tagging. RAM makes substantial step large models in computer vision, demonstrating zero-shot ability to recognize any common category with high accuracy. introduces new paradigm tagging, leveraging large-scale image-text pairs training instead of manual annotations. The development comprises four key steps. Firstly, annotation-free tags are obtained at scale through automatic text semantic parsing....
Active learning selects informative samples for annotation within budget, which has proven efficient recently on object detection. However, the widely used active detection benchmarks conduct image-level evaluation, is unrealistic in human workload estimation and biased towards crowded images. Furthermore, existing methods still perform annotation, but equally scoring all targets same image incurs waste of budget redundant labels. Having revealed above problems limitations, we introduce a...
Traditional Chinese Medicine (TCM) prescriptions are derived from the distinctive thought process and clinical experiences of medical theory. With advent artificial intelligence (AI), there is an enhanced ability to formulate these by analyzing symptom data. However, inherent sparseness herb-symptom association data still limits efficacy such predictive methods. This study introduces bipartite graph diffusion algorithm coupled with a gated recurrent self-attention mechanism for predicting...
<title>Abstract</title> The cortico-basal ganglia (BG) circuit is vital for motor control and Parkinson's disease (PD) symptoms. Globus Pallidus Internus (GPi) a principal BG output nucleus key target Deep Brain Stimulation (DBS) in PD treatment. However, the structure function of cortico-GPi have not been completely addressed. In present studies, data demonstrate direct Primary Motor Cortex (M1)-GPi pathway, bypassing classical direct, indirect, hyperdirect pathways. This M1-GPi pathway...
The inhibitory control dysfunction associated with the cognitive symptoms resulting from repetitive subconcussion (SC) is frequent. Implementing temporally resolved and likely related to dynamic interactions in functional brain networks. However, investigations of activity these networks using electroencephalography (EEG) are often limited specific frequency bands without entirely utilizing spatiotemporal rhythmic information. Therefore, we proposed an innovative framework for constructing a...
This paper presents Tag2Text, a vision language pre-training (VLP) framework, which introduces image tagging into vision-language models to guide the learning of visual-linguistic features. In contrast prior works utilize object tags either manually labeled or automatically detected with an off-the-shelf detector limited performance, our approach explicitly learns tagger using parsed from image-paired text and thus provides strong semantic guidance models. this way, Tag2Text can large-scale...
Since industrialization, global carbon dioxide (CO2) emissions have been rising substantially, playing an increasingly important role in warming and climate change. As the largest CO2 emitter, China has proposed ambitious reduction plan of peaking before 2030 achieving neutrality by 2060. Calculation inventories at regional scales (e.g., city county) great significance terms China’s policies as well national targets. However, most existing data were calculated based on fossil fuel...
Although analyzing the brain's functional and structural network has revealed that numerous brain networks are necessary to collaborate during deception, directionality of these is still unknown. This study investigated effective connectivity deception uncovers information-interaction patterns lying neural oscillations. The electroencephalography (EEG) data 40 persons honest were used create weight- directed (WDFBN). Specifically, connecting edge weight was defined based on normalized phase...