Deep Semantic Matching Research Articles

A Large Language Model-based Approach for Personalized Search Results Re-ranking in Professional Domains

Search result personalization and re-ranking in professional domains present significant challenges due to the complexity of domain-specific terminology and varying user expertise levels. This paper proposes a novel framework integrating Large Language Models (LLMs) with personalized search re-ranking for professional domains. The framework incorporates four key components: LLM-based user profile construction, professional domain knowledge encoding, cross-encoder re-ranking, and dynamic weight allocation. The user profile construction module utilizes historical interactions and professional behaviors to generate comprehensive user representations, while the domain knowledge encoding module captures specialized terminology and relationships. A cross-encoder architecture performs deep semantic matching between queries and documents, with results optimized through a dynamic weight allocation strategy. Experimental evaluation on three professional datasets (MedSearch, LegalDoc, and TechQuery) demonstrates significant improvements over existing methods, achieving 15.2% higher nDCG@10 and 12.8% better MRR compared to traditional ranking approaches. The framework maintains stable performance under varying query loads while effectively handling domain-specific terminology and user expertise variations. Ablation studies reveal the substantial impact of each component, with the LLM-based user profile construction and domain knowledge encoding contributing 7.5% and 6.3% improvements respectively. The proposed approach establishes new benchmarks for professional search systems while providing insights into effective integration of LLMs in domain-specific information retrieval.

A study of deep semantic matching in question-and-answer events in civil litigation in the environmental justice system

Abstract Information retrieval and text mining fields extensively utilize text semantic matching models. In this paper, civil litigation Q&A under the environmental justice system is taken as a specific research field, and after constructing a civil litigation Q&A system based on deep learning, two of the key techniques—question categorization and semantic matching—are selected as the main research content. Specifically, the ALBERT algorithm is used to extract word vectors, and the hidden feature vectors are obtained through BiLSTM modeling of contextual relationships and then combined with the Attention mechanism for scoring and weighting to obtain the final text-level vectors for classification so as to establish the civil litigation question classification model based on ALBERT. Then, we establish the BERT-based civil litigation question and answer matching model by sorting the set of candidate answers by semantic matching degree based on the BERT algorithm. Selected datasets and comparison algorithms are experimented with, and the analysis shows that the question classification model has a better effect than civil litigation question text classification, and the values of each index have been improved by 0.75%~3.00% on the basis of the baseline model. The MAP and MRR values (0.76~0.86) of the question-matching model are higher than those of the comparison model, verifying its superior performance in semantically assigning characters. The model proposed in this paper is more useful because it can provide civil litigation counseling to the public.

Automatic ICD-10 coding: Deep semantic matching based on analogical reasoning

BackgroundICD-10 has been widely used in statistical analysis of mortality rates and medical reimbursement. Automatic ICD-10 coding is desperately needed because manually assigning codes is expensive, time-consuming, and labor-intensive. Diagnoses described in medical records differ significantly from those used in ICD-10 classification, making it impossible for existing automatic coding techniques to perform well enough to support medical billing, resource allocation, and research requirements. Meanwhile, most of the current automatic coding approaches are oriented toward English ICD-10. This method for automatically assigning ICD-10 codes to diagnoses extracted from Chinese discharge records was provided in this paper. MethodFirst, BERT creates word representations of the two texts. Second, the context representation layer incorporates contextual information into the representation of each time step of the word representations using a bidirectional Long Short-Term Memory. Third, the matching layer compares each contextual embedding of the uncoded diagnosis record against a weighted version of all contextual character embeddings of the manually coded diagnosis record. The matching strategy is element-wise subtraction and element-wise multiplication and then through a neural network layer. Fourth, the matching vectors are combined using a one-layer convolutional neural network. A sigmoid is then used to output matching results. ResultsTo evaluate the proposed method, 1,003,558 manually coded primary diagnoses were gathered from the homepage of the discharge medical records. The experimental results showed that the proposed method outperformed popular deep semantic matching algorithms, such as DSSM, ConvNet, ESIM, and ABCNN, and demonstrated state-of-the-art results in a single text matching with an accuracy of 0.986, a precision of 0.979, a recall of 0.983, and an F1-score of 0.981. ConclusionThe automatic ICD-10 coding of Chinese diagnoses is successful when using the proposed deep semantic matching approach based on analogical reasoning.

Swings and Roundabouts: Attention-Structure Interaction Effect in Deep Semantic Matching

In the context of deep learning models for semantic matching problems, we propose a novel Multi-View Progressive Attention (MV-PA) mechanism general enough to operate on various linguistic structures of text. More importantly, we study the interaction effect between explicit linguistic structures (e.g., linear, constituency, and dependency) and implicit structures elicited by attention mechanisms . Empirical results on multiple datasets demonstrate salient patterns of substitutability between the two families of structures (explicit and implicit). Our findings not only provide intellectual foundations for the popular use of “linear LSTM + attention” architectures in NLP/QA research, but also have implications in other modalities and domains.

Combining Fact Extraction and Verification with Neural Semantic Matching Networks

The increasing concern with misinformation has stimulated research efforts on automatic fact checking. The recentlyreleased FEVER dataset introduced a benchmark factverification task in which a system is asked to verify a claim using evidential sentences from Wikipedia documents. In this paper, we present a connected system consisting of three homogeneous neural semantic matching models that conduct document retrieval, sentence selection, and claim verification jointly for fact extraction and verification. For evidence retrieval (document retrieval and sentence selection), unlike traditional vector space IR models in which queries and sources are matched in some pre-designed term vector space, we develop neural models to perform deep semantic matching from raw textual input, assuming no intermediate term representation and no access to structured external knowledge bases. We also show that Pageview frequency can also help improve the performance of evidence retrieval results, that later can be matched by using our neural semantic matching network. For claim verification, unlike previous approaches that simply feed upstream retrieved evidence and the claim to a natural language inference (NLI) model, we further enhance the NLI model by providing it with internal semantic relatedness scores (hence integrating it with the evidence retrieval modules) and ontological WordNet features. Experiments on the FEVER dataset indicate that (1) our neural semantic matching method outperforms popular TF-IDF and encoder models, by significant margins on all evidence retrieval metrics, (2) the additional relatedness score and WordNet features improve the NLI model via better semantic awareness, and (3) by formalizing all three subtasks as a similar semantic matching problem and improving on all three stages, the complete model is able to achieve the state-of-the-art results on the FEVER test set (two times greater than baseline results).1

MPSC: A Multiple-Perspective Semantics-Crossover Model for Matching Sentences

Sentence matching is crucial to many natural language processing (NLP) tasks. Generally, the degree of matching is measured from either of the two perspectives: topic-based match or semantic-based match. The former is to investigate if two sentences discuss the same topic, and the latter performs a deep level semantic matching of texts, which is currently the highlight in research. Deep semantic matching requires adequate modeling from the internal structure of the language objects as well as their interactions. To achieve this goal, this paper proposes a multiple-perspective semantics-crossover (MPSC) model for modeling the semantic-based match of two sentences. The model extracts the matching information of two sentences from the semantic interaction information generated from different angles, so as to calculate the matching degree of the two sentences. The MPSC model not only captures rich matching patterns at different levels but also acquires interactive features from different semantic angles. It can be used to address some important issues in NLP fields, such as information matching in text retrieval, question-answer matching in the Q&A system, and so on. The experimental results show that our proposed model of MPSC has better effectiveness than some popular semantic matching approaches.

Deep-MATEM: TEM query image based cross-modal retrieval for material science literature

With the rapid increasing of published material science literatures, an effective literature retrieving system is important for researchers to obtain relevant information. In this paper we propose a cross-modal material science literatures retrieval method using transmission electron microscopy(TEM) image as query information, which provide a access of using material experiment generated TEM image data to retrieve literatures. In this method, terminologies are extracted and topic distribution are inferred from text part of literatures by using LDA, and we design a multi-task Convolutional Neuron Network(CNN) mapping query TEM image to the relevant terminologies and topic distribution predictions. The ranking score is calculated from output for query image and text data. Experimental results shows our method achieves better performance than multi-label CCA, Deep Semantic Matching(Deep SM) and Modality-Specific Deep Structure(MSDS).

Cross-Modal Retrieval With CNN Visual Features: A New Baseline.

Recently, convolutional neural network (CNN) visual features have demonstrated their powerful ability as a universal representation for various recognition tasks. In this paper, cross-modal retrieval with CNN visual features is implemented with several classic methods. Specifically, off-the-shelf CNN visual features are extracted from the CNN model, which is pretrained on ImageNet with more than one million images from 1000 object categories, as a generic image representation to tackle cross-modal retrieval. To further enhance the representational ability of CNN visual features, based on the pretrained CNN model on ImageNet, a fine-tuning step is performed by using the open source Caffe CNN library for each target data set. Besides, we propose a deep semantic matching method to address the cross-modal retrieval problem with respect to samples which are annotated with one or multiple labels. Extensive experiments on five popular publicly available data sets well demonstrate the superiority of CNN visual features for cross-modal retrieval.