A5070094915- Topic Modeling
- Semantic Web and Ontologies
- Natural Language Processing Techniques
- Data Quality and Management
- Advanced Graph Neural Networks
- Multimodal Machine Learning Applications
- Text and Document Classification Technologies
- Web Data Mining and Analysis
- Domain Adaptation and Few-Shot Learning
- Biomedical Text Mining and Ontologies
- Mobile Crowdsensing and Crowdsourcing
- Expert finding and Q&A systems
- Speech and dialogue systems
- Advanced Text Analysis Techniques
- Wikis in Education and Collaboration
- Speech Recognition and Synthesis
- Data Stream Mining Techniques
- Data Mining Algorithms and Applications
- Information Retrieval and Search Behavior
- Data Management and Algorithms
- Video Surveillance and Tracking Methods
- Face and Expression Recognition
- Online Learning and Analytics
- Complex Network Analysis Techniques
- Educational Technology and Assessment
Xi'an Jiaotong University
2014-2024
In the era of big data, knowledge engineering faces fundamental challenges induced by fragmented from heterogeneous, autonomous sources with complex and evolving relationships. The representation, acquisition, inference techniques developed in 1970s 1980s, driven research development expert systems, must be updated to cope both multiple data revolution in-depth domain experts. This article presents BigKE, a framework that handles modeling online learning information sources, nonlinear fusion...
In this work, we address the challenging task of Generalized Referring Expression Comprehension (GREC). Compared to classic (REC) that focuses on single-target expressions, GREC extends scope a more practical setting by further encompassing no-target and multi-target expressions. Existing REC methods face challenges in handling complex cases encountered GREC, primarily due their fixed output limitations multi-modal representations. To these issues, propose Hierarchical Alignment-enhanced...
In this work, we address the challenging task of Generalized Referring Expression Comprehension (GREC). Compared to classic (REC) that focuses on single-target expressions, GREC extends scope a more practical setting by further encompassing no-target and multi-target expressions. Existing REC methods face challenges in handling complex cases encountered GREC, primarily due their fixed output limitations multi-modal representations. To these issues, propose Hierarchical Alignment-enhanced...
Discovering hyponym relations among domain-specific terms is a fundamental task in taxonomy learning and knowledge acquisition. However, the great diversity of various domain corpora lack labeled training sets make this very challenging for conventional methods that are based on text content. The hyperlink structure Wikipedia article pages was found to contain recurring network motifs study, indicating probability being hyperlink. Hence, novel relation extraction approach hyperlinks...
Textbook Question Answering (TQA) task requires answering questions by reasoning based on both the given diagrams and text context. There are mainly two challenges for task. First, different from natural images. Similar shapes or color blocks may express semantics there is also a large intra-topic variation diagrams. Hence, characteristics of visual semantic ambiguity variable appearance make diagram understanding more challenging. Second, text, specific education domain with terminologies...
Considerable effort has been made to increase the scale of Linked Data. However, because openness Semantic Web and ease extracting Data from semi-structured sources (e.g., Wikipedia) unstructured sources, many often provide conflicting objects for a certain predicate real-world entity. Existing methods cannot be trivially extended resolve conflicts in scale-free property. In this demonstration, we present novel system called TruthDiscover, identify truth with First, TruthDiscover leverages...
Text segmentation is a fundamental step in natural language processing (NLP) and information retrieval (IR) tasks. Most existing approaches do not explicitly take into account the facet of documents for segmentation. annotation are often addressed as separate problems, but they operate common input space. This article proposes FTS, which novel model faceted text via multitask learning (MTL). FTS models an MTL problem with annotation. employs bidirectional long short-term memory (Bi-LSTM)...
Conversational question generation aims to generate questions that depend on both context and conversation history. Conventional works utilizing deep learning have shown promising results, but heavily rely the availability of large-scale annotated conversations. In this paper, we introduce a more realistic less explored setting, Zero-shot Question Generation (ZeroCQG), which requires no human-labeled conversations for training. To solve ZeroCQG, propose multi-stage knowledge transfer...
Extracting faceted taxonomies from the Web has received increasing attention in recent years web mining community. We demonstrate this study a novel system called DFT-Extractor, which automatically constructs domain-specific Wikipedia three steps: 1) It crawls domain terms by using modified topical crawler. 2) Then it exploits classification model to extract hyponym relations with use of motif-based features. 3) Finally, taxonomy applying community detection algorithm and group heuristic...
Most community question answering (CQA) websites manage plenty of question-answer pairs (QAPs) through topic-based organizations, which may not satisfy users' fine-grained search demands. Facets topics serve as a powerful tool to navigate, refine, and group the QAPs. In this work, we propose FACM, model annotate QAPs with facets by extending convolution neural networks (CNNs) matching strategy. First, phrase information is incorporated into text representation CNNs different kernel sizes....