A5109785796- Natural Language Processing Techniques
- Topic Modeling
- Semantic Web and Ontologies
- Multimodal Machine Learning Applications
- Biological Research and Disease Studies
- Domain Adaptation and Few-Shot Learning
- Intelligent Tutoring Systems and Adaptive Learning
- Radiomics and Machine Learning in Medical Imaging
- Human Pose and Action Recognition
- Advanced Nanomaterials in Catalysis
- Swearing, Euphemism, Multilingualism
- Smart Grid Energy Management
- Translation Studies and Practices
- COVID-19 diagnosis using AI
- Second Language Acquisition and Learning
- AI in cancer detection
- Obstructive Sleep Apnea Research
- Microgrid Control and Optimization
- X-ray Diffraction in Crystallography
- Sleep and Work-Related Fatigue
- Optimal Power Flow Distribution
- Crystallization and Solubility Studies
Yangtze River Pharmaceutical Group (China)
2025
Chinese Academy of Sciences
2023
Beijing National Laboratory for Molecular Sciences
2023
Interface (United States)
2023
University of Chinese Academy of Sciences
2023
University of Science and Technology of China
2023
University of Toronto
2022
Vector Institute
2022
A conversational system is an artificial intelligence application designed to interact with users in natural language, providing accurate and contextually relevant responses. Building such systems for low-resource languages like Swahili presents significant challenges due the limited availability of large-scale training datasets. This paper proposes a Retrieval-Augmented Generation-based address these improve quality AI. The leverages fine-tuning, where models are trained on available data,...
Slang is a predominant form of informal language making flexible and extended use words that notoriously hard for natural processing systems to interpret. Existing approaches slang interpretation tend rely on context but ignore semantic extensions common in word usage. We propose semantically informed (SSI) framework considers jointly the contextual appropriateness candidate query slang. perform rigorous evaluation two large-scale online dictionaries show our approach not only achieves...
Large-scale high-quality training data is important for improving the performance of models. After trained with that has rationales (reasoning steps), models gain reasoning capability. However, dataset relatively scarce due to high annotation cost. To address this issue, we propose \textit{Self-motivated Learning} framework. The framework motivates model itself automatically generate on existing datasets. Based inherent rank from correctness across multiple rationales, learns better leading...
Through reading the documentation in context, tool-using language models can dynamically extend their capability using external tools. The cost is that we have to input lengthy every time model needs use tool, occupying window as well slowing down decoding process. Given progress general-purpose compression, soft context compression a suitable approach alleviate problem. However, when compressing tool documentation, existing methods suffer from weaknesses of key information loss...
Although the development of new energy sources such as wind and solar has alleviated demand for in China to some extent, its large-scale grid connection brought challenges safe, stable efficient operation power grid. Based on independent microgrid, this paper puts forward a coordinated optimization model storage based improved GA (Genetic Algorithm), studies method optimizing distribution hybrid system. The multi-hysteresis control strategy is adopted distribute among units reasonably. used...
Obstructive Sleep Apnea Syndrome (OSAS) is a prevalent health issue. Although the traditional polysomnography diagnostic method comprehensive, it presents several challenges in practical applications, such as need to wear multiple devices and disturbances sleep. To address these issues, this paper introduces non-intrusive of monitoring sleep apnea based on voice recognition technology, which involves real-time snoring detection. A intervention system was developed, primarily composed smart...
Abstract Humans can flexibly extend word usages across different grammatical classes, a phenomenon known as class conversion. Noun-to-verb conversion, or denominal verb (e.g., to Google cheap flight), is one of the most prevalent forms However, existing natural language processing systems are impoverished in interpreting and generating novel usages. Previous work has suggested that comprehensible if listener compute intended meaning based on shared knowledge with speaker. Here we explore...
Classification on smartphone-captured chest X-ray (CXR) photos to detect pathologies is challenging due the projective transformation caused by non-ideal camera position. Recently, various rectification methods have been proposed for different photo tasks such as document photos, license plate etc. Unfortunately, we found that none of them suitable CXR their specific type, image appearance, annotation In this paper, propose an innovative deep learning-based Projective Transformation...