We analyze the suitability of existing pre-trained transformer-based language models(PLMs)for abstractive text summarization on German technical healthcare texts.The study focuses on the multilingual capabilities of t...We analyze the suitability of existing pre-trained transformer-based language models(PLMs)for abstractive text summarization on German technical healthcare texts.The study focuses on the multilingual capabilities of these models and their ability to perform the task of abstractive text summarization in the healthcare field.The research hypothesis was that large language models could perform high-quality abstractive text summarization on German technical healthcare texts,even if the model is not specifically trained in that language.Through experiments,the research questions explore the performance of transformer language models in dealing with complex syntax constructs,the difference in performance between models trained in English and German,and the impact of translating the source text to English before conducting the summarization.We conducted an evaluation of four PLMs(GPT-3,a translation-based approach also utilizing GPT-3,a German language Model,and a domain-specific bio-medical model approach).The evaluation considered the informativeness using 3 types of metrics based on Recall-Oriented Understudy for Gisting Evaluation(ROUGE)and the quality of results which is manually evaluated considering 5 aspects.The results show that text summarization models could be used in the German healthcare domain and that domain-independent language models achieved the best results.The study proves that text summarization models can simplify the search for pre-existing German knowledge in various domains.展开更多
Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requir...Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic elements.Most existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing impacts.An ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation corpus.The conversational sentences are classified into four categories:information,question,directive,and commission.These classification label sequences are for analyzing the conversation progress and predicting the pecking order of the conversation.Ensemble of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with hyperparameters.Hyperparameter tuning approach is carried out for better performance on sentence classification.This Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation dataset.The proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer models.The proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.展开更多
Multimodal sentiment analysis is an essential area of research in artificial intelligence that combines multiple modes,such as text and image,to accurately assess sentiment.However,conventional approaches that rely on...Multimodal sentiment analysis is an essential area of research in artificial intelligence that combines multiple modes,such as text and image,to accurately assess sentiment.However,conventional approaches that rely on unimodal pre-trained models for feature extraction from each modality often overlook the intrinsic connections of semantic information between modalities.This limitation is attributed to their training on unimodal data,and necessitates the use of complex fusion mechanisms for sentiment analysis.In this study,we present a novel approach that combines a vision-language pre-trained model with a proposed multimodal contrastive learning method.Our approach harnesses the power of transfer learning by utilizing a vision-language pre-trained model to extract both visual and textual representations in a unified framework.We employ a Transformer architecture to integrate these representations,thereby enabling the capture of rich semantic infor-mation in image-text pairs.To further enhance the representation learning of these pairs,we introduce our proposed multimodal contrastive learning method,which leads to improved performance in sentiment analysis tasks.Our approach is evaluated through extensive experiments on two publicly accessible datasets,where we demonstrate its effectiveness.We achieve a significant improvement in sentiment analysis accuracy,indicating the supe-riority of our approach over existing techniques.These results highlight the potential of multimodal sentiment analysis and underscore the importance of considering the intrinsic semantic connections between modalities for accurate sentiment assessment.展开更多
We present an approach to classify medical text at a sentence level automatically.Given the inherent complexity of medical text classification,we employ adapters based on pre-trained language models to extract informa...We present an approach to classify medical text at a sentence level automatically.Given the inherent complexity of medical text classification,we employ adapters based on pre-trained language models to extract information from medical text,facilitating more accurate classification while minimizing the number of trainable parameters.Extensive experiments conducted on various datasets demonstrate the effectiveness of our approach.展开更多
Rock thin section description is an essential method for examining lithology,structure,diagenesis,and sedimentary environment,playing a pivotal role in fields such as geology,geophysics,and petroleum exploration.To ov...Rock thin section description is an essential method for examining lithology,structure,diagenesis,and sedimentary environment,playing a pivotal role in fields such as geology,geophysics,and petroleum exploration.To overcome the challenges of subjectivity,low efficiency,and high expertise requirements in describing rock thin sections,we design a multimodal mapping network,ThinGPT,which aligns the feature spaces of the contrastive language-image pre-training(CLIP)and Generative Pre-trained(GPT-2)through network training.Given the high frequency of keywords and the structured sentence patterns in thin-section descriptions,we introduce a tokenization method tailored for rock thin sections.This approach enhances GPT-2's ability to effectively encode text and produce text feature vectors.We conducted comparative experiments using ThinGPT and other models on common sedimentary rocks.The results demonstrate that ThinGPT exhibits excellent potential in generating thin-section feature descriptions of rocks.Based on the geological expert evaluation criteria proposed in this study,ThinGPT achieved a score of 1.62 on the test set.For model complexity,ThinGPT avoids heavy initial training of large language models(LLMs).This training strategy makes the model lighter and improves the efficiency of rock thin section descriptions.As an innovative application of a LLMs within a lightweight architecture for rock thin section description,ThinGPT has significant implications for intelligent geology,geophysics,and petroleum exploration.展开更多
This letter evaluates the article by Gravina et al on ChatGPT’s potential in providing medical information for inflammatory bowel disease patients.While promising,it highlights the need for advanced techniques like r...This letter evaluates the article by Gravina et al on ChatGPT’s potential in providing medical information for inflammatory bowel disease patients.While promising,it highlights the need for advanced techniques like reasoning+action and retrieval-augmented generation to improve accuracy and reliability.Emphasizing that simple question and answer testing is insufficient,it calls for more nuanced evaluation methods to truly gauge large language models’capabilities in clinical applications.展开更多
A large language model(LLM)is constructed to address the sophisticated demands of data retrieval and analysis,detailed well profiling,computation of key technical indicators,and the solutions to complex problems in re...A large language model(LLM)is constructed to address the sophisticated demands of data retrieval and analysis,detailed well profiling,computation of key technical indicators,and the solutions to complex problems in reservoir performance analysis(RPA).The LLM is constructed for RPA scenarios with incremental pre-training,fine-tuning,and functional subsystems coupling.Functional subsystem and efficient coupling methods are proposed based on named entity recognition(NER),tool invocation,and Text-to-SQL construction,all aimed at resolving pivotal challenges in developing the specific application of LLMs for RDA.This study conducted a detailed accuracy test on feature extraction models,tool classification models,data retrieval models and analysis recommendation models.The results indicate that these models have demonstrated good performance in various key aspects of reservoir dynamic analysis.The research takes some injection and production well groups in the PK3 Block of the Daqing Oilfield as an example for testing.Testing results show that our model has significant potential and practical value in assisting reservoir engineers with RDA.The research results provide a powerful support to the application of LLM in reservoir performance analysis.展开更多
To improve the accuracy and generalization of well logging curve reconstruction,this paper proposes an artificial intelligence large language model“Gaia”and conducts model evaluation experiments.By fine-tuning the p...To improve the accuracy and generalization of well logging curve reconstruction,this paper proposes an artificial intelligence large language model“Gaia”and conducts model evaluation experiments.By fine-tuning the pre-trained large language model,the Gaia significantly improved its ability in extracting sequential patterns and spatial features from well-log curves.Leveraging the adapter method for fine-tuning,this model required training only about 1/70 of its original parameters,greatly improving training efficiency.Comparative experiments,ablation experiments,and generalization experiments were designed and conducted using well-log data from 250 wells.In the comparative experiment,the Gaia model was benchmarked against cutting-edge small deep learning models and conventional large language models,demonstrating that the Gaia model reduced the mean absolute error(MAE)by at least 20%.In the ablation experiments,the synergistic effect of the Gaia model's multiple components was validated,with its MAE being at least 30%lower than that of single-component models.In the generalization experiments,the superior performance of the Gaia model in blind-well predictions was further confirmed.Compared to traditional models,the Gaia model is significantly superior in accuracy and generalization for logging curve reconstruction,fully showcasing the potential of large language models in the field of well-logging.This provides a new approach for future intelligent logging data processing.展开更多
In task-oriented dialogue systems, intent, emotion, and actions are crucial elements of user activity. Analyzing the relationships among these elements to control and manage task-oriented dialogue systems is a challen...In task-oriented dialogue systems, intent, emotion, and actions are crucial elements of user activity. Analyzing the relationships among these elements to control and manage task-oriented dialogue systems is a challenging task. However, previous work has primarily focused on the independent recognition of user intent and emotion, making it difficult to simultaneously track both aspects in the dialogue tracking module and to effectively utilize user emotions in subsequent dialogue strategies. We propose a Multi-Head Encoder Shared Model (MESM) that dynamically integrates features from emotion and intent encoders through a feature fusioner. Addressing the scarcity of datasets containing both emotion and intent labels, we designed a multi-dataset learning approach enabling the model to generate dialogue summaries encompassing both user intent and emotion. Experiments conducted on the MultiWoZ and MELD datasets demonstrate that our model effectively captures user intent and emotion, achieving extremely competitive results in dialogue state tracking tasks.展开更多
Recently, the emergence of pre-trained models(PTMs) has brought natural language processing(NLP) to a new era. In this survey, we provide a comprehensive review of PTMs for NLP. We first briefly introduce language rep...Recently, the emergence of pre-trained models(PTMs) has brought natural language processing(NLP) to a new era. In this survey, we provide a comprehensive review of PTMs for NLP. We first briefly introduce language representation learning and its research progress. Then we systematically categorize existing PTMs based on a taxonomy from four different perspectives. Next,we describe how to adapt the knowledge of PTMs to downstream tasks. Finally, we outline some potential directions of PTMs for future research. This survey is purposed to be a hands-on guide for understanding, using, and developing PTMs for various NLP tasks.展开更多
Entity alignment(EA)is crucial for knowledge fusion and integration,as it aims to match equivalent entities across different KGs.Recently,many neural-based EA methods have been proposed,focusing on developing various ...Entity alignment(EA)is crucial for knowledge fusion and integration,as it aims to match equivalent entities across different KGs.Recently,many neural-based EA methods have been proposed,focusing on developing various graph representation learning models to match entities in vector spaces.However,most real-world KGs are large-scale and contain rich structural and attribute information about entities,presenting challenges for current approaches designed primarily for small-and medium-sized KGs.To address the challenges of large-scale EA,this paper introduces a simple,effective,and scalable method based on language models.Our approach first leverages the capabilities of language models to encode entities'multi-view information into low-dimensional embeddings,identifying potential aligned entity pairs with high similarity.These candidates are then re-ranked using a global matching algorithm to produce the final alignments.Experimental results show that our method achieves state-of-the-art performance on real-world large-scale EA datasets,with superior accuracy and efficiency compared to existing methods.展开更多
Multimodal sentence summarization(MMSS)is a new yet challenging task that aims to generate a concise summary of a long sentence and its corresponding image.Although existing methods have gained promising success in MM...Multimodal sentence summarization(MMSS)is a new yet challenging task that aims to generate a concise summary of a long sentence and its corresponding image.Although existing methods have gained promising success in MMSS,they overlook the powerful generation ability of generative pre-trained language models(GPLMs),which have shown to be effective in many text generation tasks.To fill this research gap,we propose to using GPLMs to promote the performance of MMSS.Notably,adopting GPLMs to solve MMSS inevitably faces two challenges:1)What fusion strategy should we use to inject visual information into GPLMs properly?2)How to keep the GPLM′s generation ability intact to the utmost extent when the visual feature is injected into the GPLM.To address these two challenges,we propose a vision enhanced generative pre-trained language model for MMSS,dubbed as Vision-GPLM.In Vision-GPLM,we obtain features of visual and textual modalities with two separate encoders and utilize a text decoder to produce a summary.In particular,we utilize multi-head attention to fuse the features extracted from visual and textual modalities to inject the visual feature into the GPLM.Meanwhile,we train Vision-GPLM in two stages:the vision-oriented pre-training stage and fine-tuning stage.In the vision-oriented pre-training stage,we particularly train the visual encoder by the masked language model task while the other components are frozen,aiming to obtain homogeneous representations of text and image.In the fine-tuning stage,we train all the components of Vision-GPLM by the MMSS task.Extensive experiments on a public MMSS dataset verify the superiority of our model over existing baselines.展开更多
Unsupervised text simplification has attracted much attention due to the scarcity of high-quality parallel text simplification corpora. Recent an unsupervised statistical text simplification based on phrase-based mach...Unsupervised text simplification has attracted much attention due to the scarcity of high-quality parallel text simplification corpora. Recent an unsupervised statistical text simplification based on phrase-based machine translation system (UnsupPBMT) achieved good performance, which initializes the phrase tables using the similar words obtained by word embedding modeling. Since word embedding modeling only considers the relevance between words, the phrase table in UnsupPBMT contains a lot of dissimilar words. In this paper, we propose an unsupervised statistical text simplification using pre-trained language modeling BERT for initialization. Specifically, we use BERT as a general linguistic knowledge base for predicting similar words. Experimental results show that our method outperforms the state-of-the-art unsupervised text simplification methods on three benchmarks, even outperforms some supervised baselines.展开更多
Earth observations,especially satellite data,have produced a wealth of methods and results in meeting global challenges,often presented in unstructured texts such as papers or reports.Accurate extraction of satellite ...Earth observations,especially satellite data,have produced a wealth of methods and results in meeting global challenges,often presented in unstructured texts such as papers or reports.Accurate extraction of satellite and instrument entities from these unstructured texts can help to link and reuse Earth observation resources.The direct use of an existing dictionary to extract satellite and instrument entities suffers from the problem of poor matching,which leads to low recall.In this study,we present a named entity recognition model to automatically extract satellite and instrument entities from unstructured texts.Due to the lack of manually labeled data,we apply distant supervision to automatically generate labeled training data.Accordingly,we fine-tune the pre-trained language model with early stopping and a weighted cross-entropy loss function.We propose the dictionary-based self-training method to correct the incomplete annotations caused by the distant supervision method.Experiments demonstrate that our method achieves significant improvements in both precision and recall compared to dictionary matching or standard adaptation of pre-trained language models.展开更多
Open Relation Extraction(ORE)is a task of extracting semantic relations from a text document.Current ORE systems have significantly improved their efficiency in obtaining Chinese relations,when compared with conventio...Open Relation Extraction(ORE)is a task of extracting semantic relations from a text document.Current ORE systems have significantly improved their efficiency in obtaining Chinese relations,when compared with conventional systems which heavily depend on feature engineering or syntactic parsing.However,the ORE systems do not use robust neural networks such as pre-trained language models to take advantage of large-scale unstructured data effectively.In respons to this issue,a new system entitled Chinese Open Relation Extraction with Knowledge Enhancement(CORE-KE)is presented in this paper.The CORE-KE system employs a pre-trained language model(with the support of a Bidirectional Long Short-Term Memory(BiLSTM)layer and a Masked Conditional Random Field(Masked CRF)layer)on unstructured data in order to improve Chinese open relation extraction.Entity descriptions in Wikidata and additional knowledge(in terms of triple facts)extracted from Chinese ORE datasets are used to fine-tune the pre-trained language model.In addition,syntactic features are further adopted in the training stage of the CORE-KE system for knowledge enhancement.Experimental results of the CORE-KE system on two large-scale datasets of open Chinese entities and relations demonstrate that the CORE-KE system is superior to other ORE systems.The F1-scores of the CORE-KE system on the two datasets have given a relative improvement of 20.1%and 1.3%,when compared with benchmark ORE systems,respectively.The source code is available at https:/github.COm/cjwen15/CORE-KE.展开更多
Medical named entity recognition(NER)is an area in which medical named entities are recognized from medical texts,such as diseases,drugs,surgery reports,anatomical parts,and examination documents.Conventional medical ...Medical named entity recognition(NER)is an area in which medical named entities are recognized from medical texts,such as diseases,drugs,surgery reports,anatomical parts,and examination documents.Conventional medical NER methods do not make full use of un-labelled medical texts embedded in medical documents.To address this issue,we proposed a medical NER approach based on pre-trained language models and a domain dictionary.First,we constructed a medical entity dictionary by extracting medical entities from labelled medical texts and collecting medical entities from other resources,such as the YiduN4 K data set.Second,we employed this dictionary to train domain-specific pre-trained language models using un-labelled medical texts.Third,we employed a pseudo labelling mechanism in un-labelled medical texts to automatically annotate texts and create pseudo labels.Fourth,the BiLSTM-CRF sequence tagging model was used to fine-tune the pre-trained language models.Our experiments on the un-labelled medical texts,which were extracted from Chinese electronic medical records,show that the proposed NER approach enables the strict and relaxed F1 scores to be 88.7%and 95.3%,respectively.展开更多
With the rapid development of digital culture,a large number of cultural texts are presented in the form of digital and network.These texts have significant characteristics such as sparsity,real-time and non-standard ...With the rapid development of digital culture,a large number of cultural texts are presented in the form of digital and network.These texts have significant characteristics such as sparsity,real-time and non-standard expression,which bring serious challenges to traditional classification methods.In order to cope with the above problems,this paper proposes a new ASSC(ALBERT,SVD,Self-Attention and Cross-Entropy)-TextRCNN digital cultural text classification model.Based on the framework of TextRCNN,the Albert pre-training language model is introduced to improve the depth and accuracy of semantic embedding.Combined with the dual attention mechanism,the model’s ability to capture and model potential key information in short texts is strengthened.The Singular Value Decomposition(SVD)was used to replace the traditional Max pooling operation,which effectively reduced the feature loss rate and retained more key semantic information.The cross-entropy loss function was used to optimize the prediction results,making the model more robust in class distribution learning.The experimental results indicate that,in the digital cultural text classification task,as compared to the baseline model,the proposed ASSC-TextRCNN method achieves an 11.85%relative improvement in accuracy and an 11.97%relative increase in the F1 score.Meanwhile,the relative error rate decreases by 53.18%.This achievement not only validates the effectiveness and advanced nature of the proposed approach but also offers a novel technical route and methodological underpinnings for the intelligent analysis and dissemination of digital cultural texts.It holds great significance for promoting the in-depth exploration and value realization of digital culture.展开更多
Large language models(LLMs)have emerged as transformative tools with significant potential across healthcare and medicine.In clinical settings,they hold promises for tasks ranging from clinical decision support to pat...Large language models(LLMs)have emerged as transformative tools with significant potential across healthcare and medicine.In clinical settings,they hold promises for tasks ranging from clinical decision support to patient education.Advances in LLM agents further broaden their utility by enabling multimodal processing and multi-task handling in complex clinical workflows.However,evaluating the performance of LLMs in medical contexts presents unique challenges due to the high-risk nature of healthcare and the complexity of medical data.This paper provides a comprehensive overview of current evaluation practices for LLMs and LLM agents in medicine.We contributed 3 main aspects:First,we summarized data sources used in evaluations,including existing medical resources and manually designed clinical questions,offering a basis for LLM evaluation in medical settings.Second,we analyzed key medical task scenarios:closed-ended tasks,open-ended tasks,image processing tasks,and real-world multitask scenarios involving LLM agents,thereby offering guidance for further research across different medical applications.Third,we compared evaluation methods and dimensions,covering both automated metrics and human expert assessments,while addressing traditional accuracy measures alongside agent-specific dimensions,such as tool usage and reasoning capabilities.Finally,we identified key challenges and opportunities in this evolving field,emphasizing the need for continued research and interdisciplinary collaboration between healthcare professionals and computer scientists to ensure safe,ethical,and effective deployment of LLMs in clinical practice.展开更多
Recently,a novel protein language model(PLM)was published by Liang Hong group in Science Advances1,introducing PRIME(PRotein language model for Intelligent Masked pretraining and Environment prediction,Fig.1).PRIME is...Recently,a novel protein language model(PLM)was published by Liang Hong group in Science Advances1,introducing PRIME(PRotein language model for Intelligent Masked pretraining and Environment prediction,Fig.1).PRIME is a deep learning model designed to predict and improve protein stability and activity without relying on experimental mutagenesis data.This innovative approach leverages a vast dataset of 96 million protein sequences annotated with their host bacterial optimal growth temperatures(OGTs)to develop a model that effectively guides protein engineering across various applications.展开更多
This paper investigates the application of pre-trained Vision-Language Models(VLMs)for describing images from civil engineering materials and construction sites,with a focus on construction components,structural eleme...This paper investigates the application of pre-trained Vision-Language Models(VLMs)for describing images from civil engineering materials and construction sites,with a focus on construction components,structural elements,and materials.The novelty of this study lies in the investigation of VLMs for this specialized domain,which has not been previously addressed.As a case study,the paper evaluates ChatGPT-4v’s ability to serve as a descriptor tool by comparing its performance with three human descriptions(a civil engineer and two engineering interns).The contributions of this work include adapting a pre-trained VLM to civil engineering applications without additional fine-tuning and benchmarking its performance using both semantic similarity analysis(SentenceTransformers)and lexical similarity methods.Utilizing two datasets—one from a publicly available online repository and another manually collected by the authors—the study employs whole-text and sentence pair-wise similarity analyses to assess the model’s alignment with human descriptions.Results demonstrate that the best-performing model achieved an average similarity of 76%(4%standard deviation)when compared to human-generated descriptions.The analysis also reveals better performance on the publicly available dataset.展开更多
文摘We analyze the suitability of existing pre-trained transformer-based language models(PLMs)for abstractive text summarization on German technical healthcare texts.The study focuses on the multilingual capabilities of these models and their ability to perform the task of abstractive text summarization in the healthcare field.The research hypothesis was that large language models could perform high-quality abstractive text summarization on German technical healthcare texts,even if the model is not specifically trained in that language.Through experiments,the research questions explore the performance of transformer language models in dealing with complex syntax constructs,the difference in performance between models trained in English and German,and the impact of translating the source text to English before conducting the summarization.We conducted an evaluation of four PLMs(GPT-3,a translation-based approach also utilizing GPT-3,a German language Model,and a domain-specific bio-medical model approach).The evaluation considered the informativeness using 3 types of metrics based on Recall-Oriented Understudy for Gisting Evaluation(ROUGE)and the quality of results which is manually evaluated considering 5 aspects.The results show that text summarization models could be used in the German healthcare domain and that domain-independent language models achieved the best results.The study proves that text summarization models can simplify the search for pre-existing German knowledge in various domains.
文摘Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic elements.Most existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing impacts.An ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation corpus.The conversational sentences are classified into four categories:information,question,directive,and commission.These classification label sequences are for analyzing the conversation progress and predicting the pecking order of the conversation.Ensemble of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with hyperparameters.Hyperparameter tuning approach is carried out for better performance on sentence classification.This Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation dataset.The proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer models.The proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.
基金supported by Science and Technology Research Project of Jiangxi Education Department.Project Grant No.GJJ2203306.
文摘Multimodal sentiment analysis is an essential area of research in artificial intelligence that combines multiple modes,such as text and image,to accurately assess sentiment.However,conventional approaches that rely on unimodal pre-trained models for feature extraction from each modality often overlook the intrinsic connections of semantic information between modalities.This limitation is attributed to their training on unimodal data,and necessitates the use of complex fusion mechanisms for sentiment analysis.In this study,we present a novel approach that combines a vision-language pre-trained model with a proposed multimodal contrastive learning method.Our approach harnesses the power of transfer learning by utilizing a vision-language pre-trained model to extract both visual and textual representations in a unified framework.We employ a Transformer architecture to integrate these representations,thereby enabling the capture of rich semantic infor-mation in image-text pairs.To further enhance the representation learning of these pairs,we introduce our proposed multimodal contrastive learning method,which leads to improved performance in sentiment analysis tasks.Our approach is evaluated through extensive experiments on two publicly accessible datasets,where we demonstrate its effectiveness.We achieve a significant improvement in sentiment analysis accuracy,indicating the supe-riority of our approach over existing techniques.These results highlight the potential of multimodal sentiment analysis and underscore the importance of considering the intrinsic semantic connections between modalities for accurate sentiment assessment.
文摘We present an approach to classify medical text at a sentence level automatically.Given the inherent complexity of medical text classification,we employ adapters based on pre-trained language models to extract information from medical text,facilitating more accurate classification while minimizing the number of trainable parameters.Extensive experiments conducted on various datasets demonstrate the effectiveness of our approach.
基金supported by a grant from the National Natural Science Foundation of China(Grant No.42174156).
文摘Rock thin section description is an essential method for examining lithology,structure,diagenesis,and sedimentary environment,playing a pivotal role in fields such as geology,geophysics,and petroleum exploration.To overcome the challenges of subjectivity,low efficiency,and high expertise requirements in describing rock thin sections,we design a multimodal mapping network,ThinGPT,which aligns the feature spaces of the contrastive language-image pre-training(CLIP)and Generative Pre-trained(GPT-2)through network training.Given the high frequency of keywords and the structured sentence patterns in thin-section descriptions,we introduce a tokenization method tailored for rock thin sections.This approach enhances GPT-2's ability to effectively encode text and produce text feature vectors.We conducted comparative experiments using ThinGPT and other models on common sedimentary rocks.The results demonstrate that ThinGPT exhibits excellent potential in generating thin-section feature descriptions of rocks.Based on the geological expert evaluation criteria proposed in this study,ThinGPT achieved a score of 1.62 on the test set.For model complexity,ThinGPT avoids heavy initial training of large language models(LLMs).This training strategy makes the model lighter and improves the efficiency of rock thin section descriptions.As an innovative application of a LLMs within a lightweight architecture for rock thin section description,ThinGPT has significant implications for intelligent geology,geophysics,and petroleum exploration.
文摘This letter evaluates the article by Gravina et al on ChatGPT’s potential in providing medical information for inflammatory bowel disease patients.While promising,it highlights the need for advanced techniques like reasoning+action and retrieval-augmented generation to improve accuracy and reliability.Emphasizing that simple question and answer testing is insufficient,it calls for more nuanced evaluation methods to truly gauge large language models’capabilities in clinical applications.
基金Supported by the National Talent Fund of the Ministry of Science and Technology of China(20230240011)China University of Geosciences(Wuhan)Research Fund(162301192687)。
文摘A large language model(LLM)is constructed to address the sophisticated demands of data retrieval and analysis,detailed well profiling,computation of key technical indicators,and the solutions to complex problems in reservoir performance analysis(RPA).The LLM is constructed for RPA scenarios with incremental pre-training,fine-tuning,and functional subsystems coupling.Functional subsystem and efficient coupling methods are proposed based on named entity recognition(NER),tool invocation,and Text-to-SQL construction,all aimed at resolving pivotal challenges in developing the specific application of LLMs for RDA.This study conducted a detailed accuracy test on feature extraction models,tool classification models,data retrieval models and analysis recommendation models.The results indicate that these models have demonstrated good performance in various key aspects of reservoir dynamic analysis.The research takes some injection and production well groups in the PK3 Block of the Daqing Oilfield as an example for testing.Testing results show that our model has significant potential and practical value in assisting reservoir engineers with RDA.The research results provide a powerful support to the application of LLM in reservoir performance analysis.
基金Supported by the National Natural Science Foundation of China(52288101)National Key R&D Program of China(2024YFF1500600)。
文摘To improve the accuracy and generalization of well logging curve reconstruction,this paper proposes an artificial intelligence large language model“Gaia”and conducts model evaluation experiments.By fine-tuning the pre-trained large language model,the Gaia significantly improved its ability in extracting sequential patterns and spatial features from well-log curves.Leveraging the adapter method for fine-tuning,this model required training only about 1/70 of its original parameters,greatly improving training efficiency.Comparative experiments,ablation experiments,and generalization experiments were designed and conducted using well-log data from 250 wells.In the comparative experiment,the Gaia model was benchmarked against cutting-edge small deep learning models and conventional large language models,demonstrating that the Gaia model reduced the mean absolute error(MAE)by at least 20%.In the ablation experiments,the synergistic effect of the Gaia model's multiple components was validated,with its MAE being at least 30%lower than that of single-component models.In the generalization experiments,the superior performance of the Gaia model in blind-well predictions was further confirmed.Compared to traditional models,the Gaia model is significantly superior in accuracy and generalization for logging curve reconstruction,fully showcasing the potential of large language models in the field of well-logging.This provides a new approach for future intelligent logging data processing.
基金funded by the Science and Technology Foundation of Chongqing EducationCommission(GrantNo.KJQN202301153)the ScientificResearch Foundation of Chongqing University of Technology(Grant No.2021ZDZ025)the Postgraduate Innovation Foundation of Chongqing University of Technology(Grant No.gzlcx20243524).
文摘In task-oriented dialogue systems, intent, emotion, and actions are crucial elements of user activity. Analyzing the relationships among these elements to control and manage task-oriented dialogue systems is a challenging task. However, previous work has primarily focused on the independent recognition of user intent and emotion, making it difficult to simultaneously track both aspects in the dialogue tracking module and to effectively utilize user emotions in subsequent dialogue strategies. We propose a Multi-Head Encoder Shared Model (MESM) that dynamically integrates features from emotion and intent encoders through a feature fusioner. Addressing the scarcity of datasets containing both emotion and intent labels, we designed a multi-dataset learning approach enabling the model to generate dialogue summaries encompassing both user intent and emotion. Experiments conducted on the MultiWoZ and MELD datasets demonstrate that our model effectively captures user intent and emotion, achieving extremely competitive results in dialogue state tracking tasks.
基金the National Natural Science Foundation of China(Grant Nos.61751201 and 61672162)the Shanghai Municipal Science and Technology Major Project(Grant No.2018SHZDZX01)and ZJLab。
文摘Recently, the emergence of pre-trained models(PTMs) has brought natural language processing(NLP) to a new era. In this survey, we provide a comprehensive review of PTMs for NLP. We first briefly introduce language representation learning and its research progress. Then we systematically categorize existing PTMs based on a taxonomy from four different perspectives. Next,we describe how to adapt the knowledge of PTMs to downstream tasks. Finally, we outline some potential directions of PTMs for future research. This survey is purposed to be a hands-on guide for understanding, using, and developing PTMs for various NLP tasks.
基金supported by the National Natural Science Foundation of China(No.62276026)。
文摘Entity alignment(EA)is crucial for knowledge fusion and integration,as it aims to match equivalent entities across different KGs.Recently,many neural-based EA methods have been proposed,focusing on developing various graph representation learning models to match entities in vector spaces.However,most real-world KGs are large-scale and contain rich structural and attribute information about entities,presenting challenges for current approaches designed primarily for small-and medium-sized KGs.To address the challenges of large-scale EA,this paper introduces a simple,effective,and scalable method based on language models.Our approach first leverages the capabilities of language models to encode entities'multi-view information into low-dimensional embeddings,identifying potential aligned entity pairs with high similarity.These candidates are then re-ranked using a global matching algorithm to produce the final alignments.Experimental results show that our method achieves state-of-the-art performance on real-world large-scale EA datasets,with superior accuracy and efficiency compared to existing methods.
文摘Multimodal sentence summarization(MMSS)is a new yet challenging task that aims to generate a concise summary of a long sentence and its corresponding image.Although existing methods have gained promising success in MMSS,they overlook the powerful generation ability of generative pre-trained language models(GPLMs),which have shown to be effective in many text generation tasks.To fill this research gap,we propose to using GPLMs to promote the performance of MMSS.Notably,adopting GPLMs to solve MMSS inevitably faces two challenges:1)What fusion strategy should we use to inject visual information into GPLMs properly?2)How to keep the GPLM′s generation ability intact to the utmost extent when the visual feature is injected into the GPLM.To address these two challenges,we propose a vision enhanced generative pre-trained language model for MMSS,dubbed as Vision-GPLM.In Vision-GPLM,we obtain features of visual and textual modalities with two separate encoders and utilize a text decoder to produce a summary.In particular,we utilize multi-head attention to fuse the features extracted from visual and textual modalities to inject the visual feature into the GPLM.Meanwhile,we train Vision-GPLM in two stages:the vision-oriented pre-training stage and fine-tuning stage.In the vision-oriented pre-training stage,we particularly train the visual encoder by the masked language model task while the other components are frozen,aiming to obtain homogeneous representations of text and image.In the fine-tuning stage,we train all the components of Vision-GPLM by the MMSS task.Extensive experiments on a public MMSS dataset verify the superiority of our model over existing baselines.
基金supported by the National Natural Science Foundation of China(Grant Nos.62076217 and 61906060)and the Program for Changjiang Scholars and Innovative Research Team in University(PCSIRT)of the Ministry of Education,China(IRT17R32).
文摘Unsupervised text simplification has attracted much attention due to the scarcity of high-quality parallel text simplification corpora. Recent an unsupervised statistical text simplification based on phrase-based machine translation system (UnsupPBMT) achieved good performance, which initializes the phrase tables using the similar words obtained by word embedding modeling. Since word embedding modeling only considers the relevance between words, the phrase table in UnsupPBMT contains a lot of dissimilar words. In this paper, we propose an unsupervised statistical text simplification using pre-trained language modeling BERT for initialization. Specifically, we use BERT as a general linguistic knowledge base for predicting similar words. Experimental results show that our method outperforms the state-of-the-art unsupervised text simplification methods on three benchmarks, even outperforms some supervised baselines.
基金supported by the National Key Research and Development Program of China:[grant number 2019YFE0126400].
文摘Earth observations,especially satellite data,have produced a wealth of methods and results in meeting global challenges,often presented in unstructured texts such as papers or reports.Accurate extraction of satellite and instrument entities from these unstructured texts can help to link and reuse Earth observation resources.The direct use of an existing dictionary to extract satellite and instrument entities suffers from the problem of poor matching,which leads to low recall.In this study,we present a named entity recognition model to automatically extract satellite and instrument entities from unstructured texts.Due to the lack of manually labeled data,we apply distant supervision to automatically generate labeled training data.Accordingly,we fine-tune the pre-trained language model with early stopping and a weighted cross-entropy loss function.We propose the dictionary-based self-training method to correct the incomplete annotations caused by the distant supervision method.Experiments demonstrate that our method achieves significant improvements in both precision and recall compared to dictionary matching or standard adaptation of pre-trained language models.
基金the high-level university construction special project of Guangdong province,China 2019(No.5041700175)the new engineering research and practice project of the Ministry of Education,China(NO.E-RGZN20201036)。
文摘Open Relation Extraction(ORE)is a task of extracting semantic relations from a text document.Current ORE systems have significantly improved their efficiency in obtaining Chinese relations,when compared with conventional systems which heavily depend on feature engineering or syntactic parsing.However,the ORE systems do not use robust neural networks such as pre-trained language models to take advantage of large-scale unstructured data effectively.In respons to this issue,a new system entitled Chinese Open Relation Extraction with Knowledge Enhancement(CORE-KE)is presented in this paper.The CORE-KE system employs a pre-trained language model(with the support of a Bidirectional Long Short-Term Memory(BiLSTM)layer and a Masked Conditional Random Field(Masked CRF)layer)on unstructured data in order to improve Chinese open relation extraction.Entity descriptions in Wikidata and additional knowledge(in terms of triple facts)extracted from Chinese ORE datasets are used to fine-tune the pre-trained language model.In addition,syntactic features are further adopted in the training stage of the CORE-KE system for knowledge enhancement.Experimental results of the CORE-KE system on two large-scale datasets of open Chinese entities and relations demonstrate that the CORE-KE system is superior to other ORE systems.The F1-scores of the CORE-KE system on the two datasets have given a relative improvement of 20.1%and 1.3%,when compared with benchmark ORE systems,respectively.The source code is available at https:/github.COm/cjwen15/CORE-KE.
基金This work is supported in part by the Guangdong Science and Technology grant(No.2016A010101033)the Hong Kong and Macao joint research and development grant with Wuyi University(No.2019WGAH21).
文摘Medical named entity recognition(NER)is an area in which medical named entities are recognized from medical texts,such as diseases,drugs,surgery reports,anatomical parts,and examination documents.Conventional medical NER methods do not make full use of un-labelled medical texts embedded in medical documents.To address this issue,we proposed a medical NER approach based on pre-trained language models and a domain dictionary.First,we constructed a medical entity dictionary by extracting medical entities from labelled medical texts and collecting medical entities from other resources,such as the YiduN4 K data set.Second,we employed this dictionary to train domain-specific pre-trained language models using un-labelled medical texts.Third,we employed a pseudo labelling mechanism in un-labelled medical texts to automatically annotate texts and create pseudo labels.Fourth,the BiLSTM-CRF sequence tagging model was used to fine-tune the pre-trained language models.Our experiments on the un-labelled medical texts,which were extracted from Chinese electronic medical records,show that the proposed NER approach enables the strict and relaxed F1 scores to be 88.7%and 95.3%,respectively.
基金funded by China National Innovation and Entrepreneurship Project Fund Innovation Training Program(202410451009).
文摘With the rapid development of digital culture,a large number of cultural texts are presented in the form of digital and network.These texts have significant characteristics such as sparsity,real-time and non-standard expression,which bring serious challenges to traditional classification methods.In order to cope with the above problems,this paper proposes a new ASSC(ALBERT,SVD,Self-Attention and Cross-Entropy)-TextRCNN digital cultural text classification model.Based on the framework of TextRCNN,the Albert pre-training language model is introduced to improve the depth and accuracy of semantic embedding.Combined with the dual attention mechanism,the model’s ability to capture and model potential key information in short texts is strengthened.The Singular Value Decomposition(SVD)was used to replace the traditional Max pooling operation,which effectively reduced the feature loss rate and retained more key semantic information.The cross-entropy loss function was used to optimize the prediction results,making the model more robust in class distribution learning.The experimental results indicate that,in the digital cultural text classification task,as compared to the baseline model,the proposed ASSC-TextRCNN method achieves an 11.85%relative improvement in accuracy and an 11.97%relative increase in the F1 score.Meanwhile,the relative error rate decreases by 53.18%.This achievement not only validates the effectiveness and advanced nature of the proposed approach but also offers a novel technical route and methodological underpinnings for the intelligent analysis and dissemination of digital cultural texts.It holds great significance for promoting the in-depth exploration and value realization of digital culture.
基金supported by the Start-up Fund for RAPs under the Strategic Hiring Scheme(Grant No.P0048623)from HKSARthe Global STEM Professorship Scheme(Grant No.P0046113)。
文摘Large language models(LLMs)have emerged as transformative tools with significant potential across healthcare and medicine.In clinical settings,they hold promises for tasks ranging from clinical decision support to patient education.Advances in LLM agents further broaden their utility by enabling multimodal processing and multi-task handling in complex clinical workflows.However,evaluating the performance of LLMs in medical contexts presents unique challenges due to the high-risk nature of healthcare and the complexity of medical data.This paper provides a comprehensive overview of current evaluation practices for LLMs and LLM agents in medicine.We contributed 3 main aspects:First,we summarized data sources used in evaluations,including existing medical resources and manually designed clinical questions,offering a basis for LLM evaluation in medical settings.Second,we analyzed key medical task scenarios:closed-ended tasks,open-ended tasks,image processing tasks,and real-world multitask scenarios involving LLM agents,thereby offering guidance for further research across different medical applications.Third,we compared evaluation methods and dimensions,covering both automated metrics and human expert assessments,while addressing traditional accuracy measures alongside agent-specific dimensions,such as tool usage and reasoning capabilities.Finally,we identified key challenges and opportunities in this evolving field,emphasizing the need for continued research and interdisciplinary collaboration between healthcare professionals and computer scientists to ensure safe,ethical,and effective deployment of LLMs in clinical practice.
文摘Recently,a novel protein language model(PLM)was published by Liang Hong group in Science Advances1,introducing PRIME(PRotein language model for Intelligent Masked pretraining and Environment prediction,Fig.1).PRIME is a deep learning model designed to predict and improve protein stability and activity without relying on experimental mutagenesis data.This innovative approach leverages a vast dataset of 96 million protein sequences annotated with their host bacterial optimal growth temperatures(OGTs)to develop a model that effectively guides protein engineering across various applications.
文摘This paper investigates the application of pre-trained Vision-Language Models(VLMs)for describing images from civil engineering materials and construction sites,with a focus on construction components,structural elements,and materials.The novelty of this study lies in the investigation of VLMs for this specialized domain,which has not been previously addressed.As a case study,the paper evaluates ChatGPT-4v’s ability to serve as a descriptor tool by comparing its performance with three human descriptions(a civil engineer and two engineering interns).The contributions of this work include adapting a pre-trained VLM to civil engineering applications without additional fine-tuning and benchmarking its performance using both semantic similarity analysis(SentenceTransformers)and lexical similarity methods.Utilizing two datasets—one from a publicly available online repository and another manually collected by the authors—the study employs whole-text and sentence pair-wise similarity analyses to assess the model’s alignment with human descriptions.Results demonstrate that the best-performing model achieved an average similarity of 76%(4%standard deviation)when compared to human-generated descriptions.The analysis also reveals better performance on the publicly available dataset.
