A5107987282- Topic Modeling
- Natural Language Processing Techniques
- Advanced Graph Neural Networks
- Speech and dialogue systems
- Advanced Text Analysis Techniques
- Transportation and Mobility Innovations
- Transportation Planning and Optimization
- Energy, Environment, and Transportation Policies
- Multimodal Machine Learning Applications
- Urban and Freight Transport Logistics
- Advanced Image and Video Retrieval Techniques
- Visual and Cognitive Learning Processes
- Video Analysis and Summarization
- Bayesian Modeling and Causal Inference
- Memory Processes and Influences
- Text Readability and Simplification
- Intelligent Tutoring Systems and Adaptive Learning
- Text and Document Classification Technologies
- Monoclonal and Polyclonal Antibodies Research
- Advanced Measurement and Detection Methods
- Advanced Biosensing Techniques and Applications
- Gene expression and cancer classification
- Infrared Target Detection Methodologies
- Color perception and design
- Graph Theory and Algorithms
Harbin Institute of Technology
2020-2024
University of Wisconsin–Madison
2021-2022
Research into the area of multiparty dialog has grown considerably over recent years. We present Molweni dataset, a machine reading comprehension (MRC) dataset with discourse structure built dialog. Molweni’s source samples from Ubuntu Chat Corpus, including 10,000 dialogs comprising 88,303 utterances. annotate 30,066 questions on this corpus, both answerable and unanswerable questions. also uniquely contributes dependency annotations in modified Segmented Discourse Representation Theory...
Multiparty Dialogue Machine Reading Comprehension (MRC) differs from traditional MRC as models must handle the complex dialogue discourse structure, previously unconsidered in MRC. To fully exploit such structure multiparty dialogue, we present a discourse-aware graph neural network, DADgraph, which explicitly constructs using dependency links and relations. validate our model, perform experiments on Molweni corpus, large-scale dataset built over annotated with structure. Experiments show...
Peptide microarrays have emerged as a powerful technology in immunoproteomics they provide tool to measure the abundance of different antibodies patient serum samples. The high dimensionality and small sample size many experiments challenge conventional statistical approaches, including those aiming control false discovery rate (FDR). Motivated by limitations reproducibility power current methods, we advance an empirical Bayesian that computes local statistics sign when provided with data on...
Multiparty Dialogue Machine Reading Comprehension (MRC) differs from traditional MRC as models must handle the complex dialogue discourse structure, previously unconsidered in MRC. To fully exploit such structure multiparty dialogue, we present a discourse-aware graph neural network, DADgraph, which explicitly constructs using dependency links and relations. validate our model, perform experiments on Molweni corpus, large-scale dataset built over annotated with structure. Experiments show...
Multi-modal relation extraction (MRE) requires the integration of multi-modal information to identify relationships between entities. Although fine-grained correlations visual objects and textual words have potential improve cross-modal interaction, they are typically modeled implicitly hindered by modality gap. This paper introduces a novel method called relational Graph-Bridged InTeraction (GBIT). GBIT aims model into interaction process explicitly. is achieved constructing graph, which...
In this paper, we present our proposed method for the shared task of ICASSP 2023 Signal Processing Grand Challenge (SPGC). We participate in Topic Title Generation (TTG), Track 3 General Meeting Understanding and (MUG) [1] SPGC. The primary objective is to generate a title that effectively summarizes given topic segment. With constraints limited model size external dataset availability, propose as Pre-training - Distillation / Fine-tuning (PDF), which can efficiently leverage knowledge from...
The benefits of retrieval practice on learning are robust and have transferred from laboratory findings to many real-world educational settings. We report two experiments that investigated a novel technique for remembering arbitrary associations (image-word pairs), with without reward as motivator. As well typical restudy conditions, we added third condition graded in which the image cue was partially released progressive process. Experiment 1 found significant over restudy, an additional...
Recently, many information processing applications appear on the web demand of user requirement. Since text is one most popular data formats across web, how to measure similarity becomes key challenge applications. Web often used record events, especially for news. One mentions multiple while only core event decides its main topic. This should take important position when measuring similarity. For this reason, paper constructs a passage-level connection graph model relations among events...
Taxi software has become one of the most popular mobile Internet software. The best subsidy schemes from different taxi company representative for times and places are achieved by a quantitative analysis using efficacy coefficient method to ease difficult problem in this paper.
In this paper, we propose the scheme for annotating large-scale multi-party chat dialogues discourse parsing and machine comprehension. The main goal of project is to help understand dialogues. Our dataset based on Ubuntu Chat Corpus. For each dialogue, annotate structure question-answer pairs As know, first large scale corpus parsing, firstly task reading
Research into the area of multiparty dialog has grown considerably over recent years. We present Molweni dataset, a machine reading comprehension (MRC) dataset with discourse structure built dialog. Molweni's source samples from Ubuntu Chat Corpus, including 10,000 dialogs comprising 88,303 utterances. annotate 30,066 questions on this corpus, both answerable and unanswerable questions. also uniquely contributes dependency annotations in modified Segmented Discourse Representation Theory...