A5101493128- Advanced Image and Video Retrieval Techniques
- Advanced Neural Network Applications
- Multimodal Machine Learning Applications
- Smart Agriculture and AI
- Remote Sensing and Land Use
- Robotics and Sensor-Based Localization
- Plant Stress Responses and Tolerance
- Software System Performance and Reliability
- Domain Adaptation and Few-Shot Learning
- Plant Molecular Biology Research
- Advanced Data and IoT Technologies
- Groundwater and Isotope Geochemistry
- Metaheuristic Optimization Algorithms Research
- Software Reliability and Analysis Research
- Efficiency Analysis Using DEA
- Vehicular Ad Hoc Networks (VANETs)
- Dermatoglyphics and Human Traits
- Internet Traffic Analysis and Secure E-voting
- Hops Chemistry and Applications
- Artificial Immune Systems Applications
- Cancer-related molecular mechanisms research
- Cancer Research and Treatment
- IoT Networks and Protocols
- Privacy-Preserving Technologies in Data
- MicroRNA in disease regulation
Chinese Academy of Sciences
2007-2024
Institute of Geochemistry
2011-2024
University of Chinese Academy of Sciences
2006-2023
China Southern Power Grid (China)
2023
Institute of Computing Technology
2019-2023
Center for Excellence in Brain Science and Intelligence Technology
2023
Fujian Normal University
2023
China University of Mining and Technology
2019-2020
University of Maryland, College Park
2013
Institute of Software
2006-2009
microRNAs (miRNAs) play important roles in plant growth and development. Previous studies have shown that down-regulation of miR398 response to oxidative stress permits up-regulation one its target genes, CSD2 (copper/zinc superoxide dismutase), thereby helps plants cope with stress. We report here heat rapidly induces reduces transcripts genes CSD1, CCS (a gene encoding a copper chaperone for both CSD1 CSD2). Transgenic expressing miR398-resistant forms under the control their native...
Crop breeding is one of the main approaches to increase crop yield and improve quality. However, process faces challenges such as complex data, difficulties in data acquisition, low prediction accuracy, resulting efficiency long cycle. Deep learning-based a strategy that applies deep learning techniques optimize process, leading accelerated improvement, enhanced efficiency, development higher-yielding, more adaptive, disease-resistant varieties for agricultural production. This perspective...
Visual object navigation is an essential task of embodied AI, which letting the agent navigate to goal under user's demand. Previous methods often focus on single-object navigation. However, in real life, human demands are generally continuous and multiple, requiring implement multiple tasks sequence. These can be addressed by repeatedly performing previous single methods. dividing into several independent perform, without global optimization between different tasks, agents' trajectories may...
Scene recognition has been a challenging task in the field of computer vision and multimedia for long time. The current scene works often extract object features through CNN, combine these two types to obtain complementary discriminative representations. However, when categories are visually similar, might lack discriminations. Therefore, it may be debatable consider only features. In contrast existing works, this paper, we discuss discrimination attributes local regions utilize as We visual...
Abstract--- Exploiting the spatial structure in scene images is a key research direction for recognition. Due to large intra-class structural diversity, building and modeling flexible layout adapt various image characteristics challenge. Existing methods recognition either focus on predefined grids or rely learned prototypes, which all have limited representative ability. In this paper, we propose Prototype-agnostic Scene Layout (PaSL) construction method build each without conforming any...
In the past few years, prediction models have shown remarkable performance in most biological correlation tasks. These tasks traditionally use a fixed dataset, and model, once trained, is deployed as is. often encounter training issues such sensitivity to hyperparameter tuning "catastrophic forgetting" when adding new data. However, with development of biomedicine accumulation data, predictive are required face challenge adapting change. To this end, we propose computational approach based...
Scene images are usually composed of foreground and background regional contents. Some existing methods propose to extract contents with dense grids or objectness region proposals. However, may split the object into several discrete parts, learning semantic ambiguity for patches. The focus on particular objects but only pay attention do not exploit that is key scene recognition. In contrast, we a novel recognition framework amorphous detection context modeling. proposed framework,...
Recognizing visual categories from semantic descriptions is a promising way to extend the capability of classifier beyond concepts represented in training data (i.e. seen categories). This problem addressed by (generalized) zero-shot learning methods (GZSL), which leverage that connect them (e.g. label embedding, attributes). Conventional GZSL are designed mostly for object recognition. In this paper we focus on scene recognition, more challenging setting with hundreds where their...
This paper presents a model of Particle Swarm Optimization with Escape Velocity (EVPSO) in order to overcome premature convergence the basic (PSO). The EVPSO equips particles escape velocity avoid them trapping into local minima and increase diversity population. A simulation study shows that outperforms PSO, especially for high dimension function. facilitates solving multi-modal optimization problems.
The goal of few-shot image recognition is to classify different categories with only one or a few training samples. Previous works learning mainly focus on simple images, such as object character images. Those usually use convolutional neural network (CNN) learn the global representations from tasks, which are then adapted novel tasks. However, there many more abstract and complex images in real world, scene consisting entities flexible spatial relations among them. In cases, features can...
Productivity is a critical performance index of process resources. As successive history productivity data tends to be auto-correlated, time series prediction method based on auto-regressive integrated moving average (ARIMA) model was introduced into software by Humphrey et al. In this paper, variant their named ARIMAmmse proposed. This formulates the ARIMA parameter estimation issue as minimum mean square error (MMSE) constrained optimization problem. The used describe constraints problem,...