Related references
Note: Only part of the references are listed.
Article
Computer Science, Hardware & Architecture
Youhui Zuo et al.
Summary: Chinese semantic textual matching is a challenging task in NLP, and accurately capturing both intra-text and inter-text features is crucial. Existing methods usually only utilize contrastive learning at a single perspective, leading to suboptimal performance. To address this, we propose a dual-stage contrastive learning framework (DuCL) for Chinese textual matching, which incorporates a block-enhanced interaction module to generate a semantic matching representation. Experimental results on real-world datasets demonstrate the superiority of our method over representative and state-of-the-art methods.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Computer Science, Information Systems
Jiang Shen et al.
Summary: Using AI technology to automatically match Q&A pairs on online health platforms can improve doctor-patient interaction efficiency. This paper proposes a model named MKGA-DM-NN, which leverages medical knowledge graph (KG) entities, graph embedding technology, and doctors' historical Q&A records to improve the accuracy of Q&A matching. Experimental results show that our model outperforms baseline models, achieving 4.4% higher accuracy and 13.53% lower cost-sensitive error. Adding doctor modeling module improves accuracy by 8.72%, while adding medical KG module reduces cost-sensitive error by 17.27%.
INFORMATION PROCESSING & MANAGEMENT
(2023)
Article
Computer Science, Information Systems
Qin Zhang et al.
Summary: Humans have the ability to reason and find answers from multiple sources. In the field of Multiple Choice Question Answering (MCQA), knowledge graphs can mimic human reasoning by providing subgraphs based on different question-answer combinations. However, current research lacks the ability for joint reasoning among all answer candidates.
INFORMATION PROCESSING & MANAGEMENT
(2023)
Article
Computer Science, Information Systems
Zongxi Li et al.
Summary: Researchers have found that emotion is not limited to one category in emotion-relevant classification tasks, and multiple emotions can exist together in a sentence. Recent studies have focused on using distribution or grayscale labels to enhance the classification model, providing additional information on the intensity of emotions and their correlations. This approach has been effective in overcoming overfitting and improving model robustness. However, it can also reduce the model's discriminative ability within similar emotion categories.
INFORMATION PROCESSING & MANAGEMENT
(2023)
Article
Computer Science, Artificial Intelligence
Xiao Pu et al.
Summary: This study proposes a method that integrates external lexical knowledge to improve text matching by modeling the senses of potentially ambiguous words. A lightweight word sense disambiguation (WSD) model based on BERT and WordNet is designed and integrated into a matching mechanism. Experimental results on three matching-based tasks show that the sense knowledge-enhanced matching mechanism outperforms BERT-based baselines and other recent approaches.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Construction & Building Technology
Lang-Tao Wu et al.
Summary: This paper proposes a rule-based approach for MEP information extraction and verifies its feasibility and efficiency through experiments.
AUTOMATION IN CONSTRUCTION
(2022)
Article
Computer Science, Information Systems
Dengwen Lin et al.
Summary: Passage ranking is crucial in information retrieval and question answering, and pre-trained language models like BERT have been shown to improve performance. However, these models can be easily fooled by overlapping but irrelevant passages. To address this issue, a self-matching attention-pooling mechanism (SMAP) and a hybrid passage ranking architecture called BERT-SMAP have been proposed to better identify distracting passages.
INFORMATION PROCESSING & MANAGEMENT
(2022)
Article
Computer Science, Information Systems
Yintao Yang et al.
Summary: CGA2TC is a new graph-based model for text classification that combines contrastive learning and adaptive augmentation strategy to obtain more robust node representation. It constructs a text graph using word co-occurrence and document word relationships and designs an augmentation strategy to solve the noise problem and preserve essential structures. The model handles labeled and unlabeled nodes differently and adopts random sampling to reduce resource consumption. Experimental results demonstrate the effectiveness of CGA2TC in text classification tasks.
INFORMATION PROCESSING & MANAGEMENT
(2022)
Article
Computer Science, Information Systems
Chunli Xiang et al.
Summary: This paper studies the methodology of predicting bullish or bearish sentiments in the financial domain. It proposes a novel semantic and syntactic enhanced neural model to enhance the accuracy of sentiment prediction by capturing contextual information and aggregating features using a graph convolutional network.
INFORMATION PROCESSING & MANAGEMENT
(2022)
Article
Computer Science, Information Systems
Chunli Xiang et al.
Summary: This paper studies the methodology of inferring bullish or bearish sentiments in the financial domain and proposes a novel neural model to tackle this task.
INFORMATION PROCESSING & MANAGEMENT
(2022)
Article
Computer Science, Information Systems
Yintao Yang et al.
Summary: This study introduces a new graph-based text classification model CGA2TC, which utilizes contrastive learning and adaptive augmentation strategy for more robust node representation. By exploring word co-occurrence and document word relationships to construct a text graph, diverse augmentation strategies are employed to address noise issues. Consistency training is applied on labeled and unlabeled nodes, demonstrating the effectiveness of the model in text classification tasks.
INFORMATION PROCESSING & MANAGEMENT
(2022)
Article
Computer Science, Information Systems
Chuanming Yu et al.
Summary: A novel Deep Interactive Text Matching (DITM) model is proposed in this study, which effectively captures the interactive information between text pairs and has high generalization ability among different tasks.
INFORMATION PROCESSING & MANAGEMENT
(2021)
Article
Computer Science, Interdisciplinary Applications
Qi Jia et al.
Summary: In this study, a novel two-step approach using hierarchical semantic information and an attention mechanism was proposed to tackle the challenges of traditional Chinese medicine symptom normalization. The approach demonstrated superior performance compared to other baselines, showing promise in this research direction.
JOURNAL OF BIOMEDICAL INFORMATICS
(2021)
Review
Computer Science, Artificial Intelligence
Zhaoyang Niu et al.
Summary: This paper provides an overview of state-of-the-art attention models and defines a unified model suitable for most attention structures. It describes in detail each step of the attention mechanism implemented in the model and classifies existing attention models based on four criteria. Additionally, it summarizes the use of attention mechanisms in network architectures and typical applications.
Article
Computer Science, Theory & Methods
Connor Shorten et al.
Summary: In this paper, we discuss how Data Augmentation can aid in the development of Natural Language Processing, covering topics such as the motivation and specific frameworks of Data Augmentation, promising ideas yet to be tested, tools for implementation, and interesting topics around Data Augmentation.
JOURNAL OF BIG DATA
(2021)
Article
Computer Science, Information Systems
Junmei Wang et al.
INFORMATION PROCESSING & MANAGEMENT
(2020)