A5050532724- Topic Modeling
- Information Retrieval and Search Behavior
- Advanced Text Analysis Techniques
- Natural Language Processing Techniques
- Sentiment Analysis and Opinion Mining
- Advanced Image and Video Retrieval Techniques
- Semantic Web and Ontologies
- Web Data Mining and Analysis
- Image Retrieval and Classification Techniques
- Text and Document Classification Technologies
- Data Management and Algorithms
- Recommender Systems and Techniques
- Expert finding and Q&A systems
- Emotion and Mood Recognition
- Multimodal Machine Learning Applications
- EEG and Brain-Computer Interfaces
- Neural Networks and Applications
- Advanced Database Systems and Queries
- Advancements in Battery Materials
- Data Quality and Management
- Video Analysis and Summarization
- Logic, Reasoning, and Knowledge
- Advanced Battery Materials and Technologies
- Algorithms and Data Compression
- Quantum Computing Algorithms and Architecture
Beijing Institute of Technology
2018-2025
First Affiliated Hospital of Xi'an Jiaotong University
2020-2025
Tianjin University of Technology
2011-2024
The Open University
2015-2024
Northeast Forestry University
2024
Shanghai Electric (China)
2023-2024
Beijing Open University
2023
Huzhou University
2023
Chinese University of Hong Kong, Shenzhen
2023
Tianjin University
2012-2022
Chen Zhang, Qiuchi Li, Dawei Song. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint (EMNLP-IJCNLP). 2019.
Recognizing cross-subject emotions based on brain imaging data, e.g., EEG, has always been difficult due to the poor generalizability of features across subjects. Thus, systematically exploring ability different EEG identify emotional information subjects is crucial. Prior related work explored this question only one or two kinds features, and findings conclusions have presented. In work, we aim at a more comprehensive investigation with wider range feature types, including 18 linear...
Automatic emotion recognition based on multi-channel neurophysiological signals, as a challenging pattern task, is becoming an important computer-aided method for emotional disorder diagnoses in neurology and psychiatry. Traditional approaches require designing extracting range of features from single or multiple channel signals extensive domain knowledge. This may be obstacle non-domain experts. Moreover, traditional feature fusion can not fully utilize correlation information between...
Language Modeling (LM) has been successfully applied to Information Retrieval (IR). However, most of the existing LM approaches only rely on term occurrences in documents, queries and document collections. In traditional unigram based models, terms (or words) are usually considered be independent. some recent studies, dependence models have proposed incorporate relationships into LM, so that links can created between words same sentence, (e.g. synonymy) used expand model. this study, we...
With the widespread applications of electronic learning (e-Learning) technologies to education at all levels, increasing number online educational resources and messages are generated from corresponding e-Learning environments. Nevertheless, it is quite difficult, if not totally impossible, for instructors read through analyze predict progress their students on fly. The main contribution this paper illustration a novel concept map generation mechanism which underpinned by fuzzy domain...
The state-of-the-art Aspect-based Sentiment Analysis (ABSA) approaches are mainly based on either detecting aspect terms and their corresponding sentiment polarities, or co-extracting opinion terms. However, the extraction of aspect-sentiment pairs lacks as a reference, while co-extraction would not lead to meaningful without determining dependencies. To address issue, we present novel view ABSA an triplet task, propose multi-task learning framework jointly extract terms, simultaneously...
Sarcasm detection in conversation, a theoretically and practically challenging artificial intelligence task, aims to discover elusively ironic, contemptuous, metaphoric information implied daily conversations. Most of the recent approaches sarcasm have neglected intrinsic vagueness uncertainty human language emotional expression understanding. To address this gap, we propose complex-valued fuzzy network by leveraging mathematical formalisms quantum theory logic. In particular, target...
Sentiment and emotion, which correspond to long-term short-lived human feelings, are closely linked each other, leading the fact that sentiment analysis emotion recognition also two interdependent tasks in natural language processing (NLP). One task often leverages shared knowledge from another performs better when solved a joint learning paradigm. Conversational context dependency, multi-modal interaction, multi-task correlation three key factors contribute this However, none of recent...
Online product reviews play an important role in E-commerce websites because most customers read and rely on them when making purchases. For the sake of profit or reputation, review spammers deliberately write fake to promote demote target products, some even fraudulently work groups try control sentiment about a product. To detect such spammer groups, previous exploits frequent itemset mining (FIM) generate candidate which can only find tightly coupled i.e. each reviewer group every In this...
Language Modeling (LM) is a fundamental research topic in range of areas. Recently, inspired by quantum theory, novel Quantum Model (QLM) has been proposed for Information Retrieval (IR). In this paper, we aim to broaden the theoretical and practical basis QLM. We develop Neural Network based Quantum-like (NNQLM) apply it Question Answering. Specifically, on word embeddings, design new density matrix, which represents sentence (e.g., question or an answer) encodes mixture semantic subspaces....
Conversational sentiment analysis is an emerging, yet challenging Artificial Intelligence (AI) subtask. It aims to discover the affective state of each participant in a conversation. There exists wealth interaction information that affects speakers. However, existing approaches are insufficient dealing with this task due ignoring interactions and dependency relationships between utterances. In paper, we aim address issue by modeling intrautterance inter-utterance dynamics. We propose...
Robust cross-subject emotion recognition based on multichannel EEG has always been a hard work. In this work, we hypothesize there exists default brain variables across subjects in emotional processes. Hence, the states of latent that related to processing must contribute building robust models. Specifically, propose utilize unsupervised deep generative model (e.g., variational autoencoder), determine factors from EEG. Through sequence modeling method, examine performance learnt factors. The...
Latest development of neural models has connected the encoder and decoder through a self-attention mechanism. In particular, Transformer, which is solely based on self-attention, led to breakthroughs in Natural Language Processing (NLP) tasks. However, multi-head attention mechanism, as key component limits effective deployment model resource-limited setting. this paper, ideas tensor decomposition parameters sharing, we propose novel (namely Multi-linear attention) with Block-Term Tensor...
3-D CAD models are an important digital resource in the manufacturing industry. model retrieval has become a key technology product lifecycle management enabling reuse of existing design data. In this paper, we propose new method to retrieve based on 2-D pen-based sketch inputs. Sketching is common and convenient for communicating intent during early stages design, e.g., conceptual design. However, converting sketched information into precise engineering cumbersome, much effort can be...