A5101584046- Topic Modeling
- Natural Language Processing Techniques
- Text and Document Classification Technologies
- Domain Adaptation and Few-Shot Learning
- Advanced Text Analysis Techniques
- Robotics and Automated Systems
- Video Surveillance and Tracking Methods
- Biometric Identification and Security
- Wireless Signal Modulation Classification
- Advanced Image and Video Retrieval Techniques
- Algorithms and Data Compression
- Advanced Measurement and Detection Methods
- Advanced Graph Neural Networks
- Neural Networks and Applications
- Biomedical Text Mining and Ontologies
- DNA and Biological Computing
Shanghai Maritime University
2024
Donghua University
2020-2021
Yunnan University
2013
Relation classification (RC) task is one of fundamental tasks information extraction, aiming to detect the relation between entity pairs in unstructured natural language text and generate structured data form entity-relation triple. Although distant supervision methods can effectively alleviate problem lack training supervised learning, they also introduce noise into still cannot fundamentally solve long-tail distribution instances. In order enable neural network learn new knowledge through...
Relation extraction is an important task of information extraction. Most existing methods Chinese language relation are based on word input. They highly dependent the quality segmentation and suffer from ambiguity polysemic words. Therefore, a multi-feature fusion model presented basis character input, which integrates character-level features, word-level features entity sense into deep neural network models. Specifically, to alleviate polysemy, introduced as external knowledge provide...
The goal of relationship classification (RC) is to predict the semantic between two entities in a given sentence. With advent deep learning and pretrained language models, RC research has progressed by leaps bounds. However, current studies are focused mainly on predicting relationships from predefined set. How recognize unseen remains challenge, which also known as zero-shot (ZSRC) task. Some ZSRC-related methods directly map categories numerical indices, constraining model's ability...
With an exponential explosive growth of various digital text information, it is challenging to efficiently obtain specific knowledge from massive unstructured information. As one basic task for natural language processing (NLP), relation extraction aims extract the semantic between entity pairs based on given text. To avoid manual labeling datasets, distant supervision (DSRE) has been widely used, aiming utilize base automatically annotate datasets. Unfortunately, this method heavily suffers...
In this paper, we proposed an optimized real-time hybrid cooperative multi-camera tracking system for large-scale au-tomate surveillance based on embedded smart cameras including stationary and moving pan/tilt/zoom (PTZ) with TI DSP TMS320DM6446 intelligent visual analysis. Firstly, the overlapping areas projection relations between adjacent cameras' field of view (FOV) is calculated. Based FOV ob-tained information each single camera, a homography target handover procedure done long-term...
Recently, locality sensitive hashing (LSH) was shown to be effective for MIPS and several algorithms including $L_2$-ALSH, Sign-ALSH Simple-LSH have been proposed. In this paper, we introduce the norm-range partition technique, which partitions original dataset into sub-datasets containing items with similar 2-norms builds hash index independently each sub-dataset. We prove that reduces query processing complexity all existing LSH based under mild conditions. The key performance improvement...
Relation classification (RC) task is one of fundamental tasks information extraction, aiming to detect the relation between entity pairs in unstructured natural language text and generate structured data form entity-relation triple. Although distant supervision methods can effectively alleviate problem lack training supervised learning, they also introduce noise into data, still cannot fundamentally solve long-tail distribution instances. In order enable neural network learn new knowledge...