A5109012722- Domain Adaptation and Few-Shot Learning
- Multimodal Machine Learning Applications
- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Nuclear Physics and Applications
- Nuclear reactor physics and engineering
- Neuroscience and Neuropharmacology Research
- Human Pose and Action Recognition
- Face and Expression Recognition
- Precipitation Measurement and Analysis
- Anomaly Detection Techniques and Applications
- Memory and Neural Mechanisms
- Advanced Graph Neural Networks
- Machine Learning and ELM
- Optical Network Technologies
- Topic Modeling
- Cancer-related molecular mechanisms research
- Flood Risk Assessment and Management
- Photoreceptor and optogenetics research
- EEG and Brain-Computer Interfaces
- Complex Network Analysis Techniques
- Neural dynamics and brain function
- Advanced Neural Network Applications
- Sparse and Compressive Sensing Techniques
- Soil Moisture and Remote Sensing
Chinese Academy of Sciences
2015-2025
Institute of Automation
2006-2025
Shandong Institute of Automation
2019-2025
Beijing Academy of Artificial Intelligence
2023
University of Chinese Academy of Sciences
2018-2023
Capital University
2023
Capital Medical University
2023
Harbin University of Science and Technology
2023
Xi'an Jiaotong University
2018-2022
Air Force Medical University
2022
The ability of animals to respond life-threatening stimuli is essential for survival. Although vision provides one the major sensory inputs detecting threats across animal species, circuitry underlying defensive responses visual remains poorly defined. Here, we investigate innate behaviours elicited by predator-like in mice. Our results demonstrate that neurons superior colliculus (SC) are a variety acute and persistent overhead looming stimuli. Optogenetic mapping revealed SC projections...
This paper presents a novel deep learning framework for the inter-patient electrocardiogram (ECG) heartbeat classification. A symbolization approach especially designed ECG is introduced, which can jointly represent morphology and rhythm of alleviate influence variation through baseline correction. The symbolic representation used by multi-perspective convolutional neural network (MPCNN) to learn features automatically classify heartbeat. We evaluate our method detection supraventricular...
Abstract The complex relationship between specific hippocampal oscillation frequency deficit and cognitive dysfunction in the ischemic brain is unclear. Here, using a mouse two-vessel occlusion (2VO) cerebral ischemia model, we show that visual stimulation with 40 Hz light flicker drove CA1 slow gamma restored 2VO-induced reduction power theta-low phase-amplitude coupling, but not those of high gamma. Low lights at 30 Hz, 50 10 80 arrhythmic light, were protective against degenerating...
Ship detection is a crucial but challenging task in optical remote sensing images. Recently, thanks to the emergence of deep neural networks, significant progress has been made ship detection. However, there are still two issues that must be addressed: 1) The high-resolution images may confuse background with ship, leading more false alarms during detection; 2) detector receives fewer positive samples due sparse and uneven distribution ships In this paper, we innovatively propose employing...
Low-rank self-representation based subspace learning has confirmed its great effectiveness in a broad range of applications. Nevertheless, existing studies mainly focus on exploring the global linear structure, and cannot commendably handle case where samples approximately (i.e., contain data errors) lie several more general affine subspaces. To overcome this drawback, paper, we innovatively propose to introduce nonnegative constraints into low-rank learning. While simple enough, provide...
Despite recent progress in semi-supervised learning (SSL), its scalability remains limited realistic scenarios where unseen classes may appear the unlabeled data. To address this challenge, open-world SSL (OWSSL) is proposed years and attracts much attention. One core difficulty OWSSL to enhance representative ability for samples, especially those novel classes. More recently, several works introduce contrastive into achieve impressive performance. However, they mainly focus on conducting...
Multiview subspace clustering (MSC) has attracted growing attention due to the extensive value in various applications, such as natural language processing, face recognition, and time-series analysis. In this article, we are devoted address two crucial issues MSC: 1) high computational cost 2) cumbersome multistage clustering. Existing MSC approaches, including tensor singular decomposition (t-SVD)-MSC that achieved promising performance, generally utilize dataset itself dictionary regard...
Multiview clustering has become a research hotspot in recent years due to its excellent capability of heterogeneous data fusion. Although great deal related works appeared one after another, most them generally overlook the potentials prior knowledge utilization and progressive sample learning, resulting unsatisfactory performance real-world applications. To with aforementioned drawbacks, this article, we propose semisupervised representation learning approach for deep multiview (namely,...
Time series clustering has attracted growing attention due to the abundant data accessible and extensive value in various applications. The unique characteristics of time series, including high-dimension, warping, integration multiple elastic measures, pose challenges for present algorithms, most which take into account only part these difficulties. In this paper, we make an effort simultaneously address all aforementioned issues under a unified kernels (MKC) framework. Specifically, first...
Multi-view subspace clustering is an effective method to partition data into their corresponding categories. Nevertheless, existing multi-view approaches generally operate in a purely unsupervised manner, while ignoring the valuable weakly supervised information that can be readily obtained many practical applications. In this paper, we consider form of sample pair constraints, and devote promoting performance with aid such prior knowledge. To achieve goal, inspired by intrinsic block...
Epilepsy, one of the world's most common neurological diseases, impacts over 1% global population. Accurate early prediction epileptic seizure has a great influence on patients' lives and attracted extensive attention. However, existing methods do not fully consider complexity multi-channel electroencephalogram (EEG) signal, which is measurement for seizures. In this letter, we propose Dynamic Functional Connectivity neural Network (DynFCNet) prediction. The proposed DynFCNet can discover...
This article presents a tensor multi-task model for person re-identification (Re-ID). Due to discrepancy among cameras, our approach regards Re-ID from multiple cameras as different but related classification tasks, each task corresponding specific camera. In task, we distinguish the identity one-vs-all linear problem, where one classifier is associated with person. By constructing all classifiers into task-specific projection matrix, proposed method could utilize matrices form structure,...