In the field of natural language processing(NLP),there have been various pre-training language models in recent years,with question answering systems gaining significant attention.However,as algorithms,data,and comput...In the field of natural language processing(NLP),there have been various pre-training language models in recent years,with question answering systems gaining significant attention.However,as algorithms,data,and computing power advance,the issue of increasingly larger models and a growing number of parameters has surfaced.Consequently,model training has become more costly and less efficient.To enhance the efficiency and accuracy of the training process while reducing themodel volume,this paper proposes a first-order pruningmodel PAL-BERT based on the ALBERT model according to the characteristics of question-answering(QA)system and language model.Firstly,a first-order network pruning method based on the ALBERT model is designed,and the PAL-BERT model is formed.Then,the parameter optimization strategy of the PAL-BERT model is formulated,and the Mish function was used as an activation function instead of ReLU to improve the performance.Finally,after comparison experiments with traditional deep learning models TextCNN and BiLSTM,it is confirmed that PALBERT is a pruning model compression method that can significantly reduce training time and optimize training efficiency.Compared with traditional models,PAL-BERT significantly improves the NLP task’s performance.展开更多
Recent advancements in natural language processing have given rise to numerous pre-training language models in question-answering systems.However,with the constant evolution of algorithms,data,and computing power,the ...Recent advancements in natural language processing have given rise to numerous pre-training language models in question-answering systems.However,with the constant evolution of algorithms,data,and computing power,the increasing size and complexity of these models have led to increased training costs and reduced efficiency.This study aims to minimize the inference time of such models while maintaining computational performance.It also proposes a novel Distillation model for PAL-BERT(DPAL-BERT),specifically,employs knowledge distillation,using the PAL-BERT model as the teacher model to train two student models:DPAL-BERT-Bi and DPAL-BERTC.This research enhances the dataset through techniques such as masking,replacement,and n-gram sampling to optimize knowledge transfer.The experimental results showed that the distilled models greatly outperform models trained from scratch.In addition,although the distilled models exhibit a slight decrease in performance compared to PAL-BERT,they significantly reduce inference time to just 0.25%of the original.This demonstrates the effectiveness of the proposed approach in balancing model performance and efficiency.展开更多
The weapon and equipment operational requirement analysis(WEORA) is a necessary condition to win a future war,among which the acquisition of knowledge about weapons and equipment is a great challenge. The main challen...The weapon and equipment operational requirement analysis(WEORA) is a necessary condition to win a future war,among which the acquisition of knowledge about weapons and equipment is a great challenge. The main challenge is that the existing weapons and equipment data fails to carry out structured knowledge representation, and knowledge navigation based on natural language cannot efficiently support the WEORA. To solve above problem, this research proposes a method based on question answering(QA) of weapons and equipment knowledge graph(WEKG) to construct and navigate the knowledge related to weapons and equipment in the WEORA. This method firstly constructs the WEKG, and builds a neutral network-based QA system over the WEKG by means of semantic parsing for knowledge navigation. Finally, the method is evaluated and a chatbot on the QA system is developed for the WEORA. Our proposed method has good performance in the accuracy and efficiency of searching target knowledge, and can well assist the WEORA.展开更多
Background External knowledge representations play an essential role in knowledge-based visual question and answering to better understand complex scenarios in the open world.Recent entity-relationship embedding appro...Background External knowledge representations play an essential role in knowledge-based visual question and answering to better understand complex scenarios in the open world.Recent entity-relationship embedding approaches are deficient in representing some complex relations,resulting in a lack of topic-related knowledge and redundancy in topic-irrelevant information.Methods To this end,we propose MKEAH:Multimodal Knowledge Extraction and Accumulation on Hyperplanes.To ensure that the lengths of the feature vectors projected onto the hyperplane compare equally and to filter out sufficient topic-irrelevant information,two losses are proposed to learn the triplet representations from the complementary views:range loss and orthogonal loss.To interpret the capability of extracting topic-related knowledge,we present the Topic Similarity(TS)between topic and entity-relations.Results Experimental results demonstrate the effectiveness of hyperplane embedding for knowledge representation in knowledge-based visual question answering.Our model outperformed state-of-the-art methods by 2.12%and 3.24%on two challenging knowledge-request datasets:OK-VQA and KRVQA,respectively.Conclusions The obvious advantages of our model in TS show that using hyperplane embedding to represent multimodal knowledge can improve its ability to extract topic-related knowledge.展开更多
ExpertRecommendation(ER)aims to identify domain experts with high expertise and willingness to provide answers to questions in Community Question Answering(CQA)web services.How to model questions and users in the hete...ExpertRecommendation(ER)aims to identify domain experts with high expertise and willingness to provide answers to questions in Community Question Answering(CQA)web services.How to model questions and users in the heterogeneous content network is critical to this task.Most traditional methods focus on modeling questions and users based on the textual content left in the community while ignoring the structural properties of heterogeneous CQA networks and always suffering from textual data sparsity issues.Recent approaches take advantage of structural proximities between nodes and attempt to fuse the textual content of nodes for modeling.However,they often fail to distinguish the nodes’personalized preferences and only consider the textual content of a part of the nodes in network embedding learning,while ignoring the semantic relevance of nodes.In this paper,we propose a novel framework that jointly considers the structural proximity relations and textual semantic relevance to model users and questions more comprehensively.Specifically,we learn topology-based embeddings through a hierarchical attentive network learning strategy,in which the proximity information and the personalized preference of nodes are encoded and preserved.Meanwhile,we utilize the node’s textual content and the text correlation between adjacent nodes to build the content-based embedding through a meta-context-aware skip-gram model.In addition,the user’s relative answer quality is incorporated to promote the ranking performance.Experimental results show that our proposed framework consistently and significantly outperforms the state-of-the-art baselines on three real-world datasets by taking the deep semantic understanding and structural feature learning together.The performance of the proposed work is analyzed in terms of MRR,P@K,and MAP and is proven to be more advanced than the existing methodologies.展开更多
Recently,pre-trained language representation models such as bidirec-tional encoder representations from transformers(BERT)have been performing well in commonsense question answering(CSQA).However,there is a problem th...Recently,pre-trained language representation models such as bidirec-tional encoder representations from transformers(BERT)have been performing well in commonsense question answering(CSQA).However,there is a problem that the models do not directly use explicit information of knowledge sources existing outside.To augment this,additional methods such as knowledge-aware graph network(KagNet)and multi-hop graph relation network(MHGRN)have been proposed.In this study,we propose to use the latest pre-trained language model a lite bidirectional encoder representations from transformers(ALBERT)with knowledge graph information extraction technique.We also propose to applying the novel method,schema graph expansion to recent language models.Then,we analyze the effect of applying knowledge graph-based knowledge extraction techniques to recent pre-trained language models and confirm that schema graph expansion is effective in some extent.Furthermore,we show that our proposed model can achieve better performance than existing KagNet and MHGRN models in CommonsenseQA dataset.展开更多
Analyzing Research and Development(R&D)trends is important because it can influence future decisions regarding R&D direction.In typical trend analysis,topic or technology taxonomies are employed to compute the...Analyzing Research and Development(R&D)trends is important because it can influence future decisions regarding R&D direction.In typical trend analysis,topic or technology taxonomies are employed to compute the popularities of the topics or codes over time.Although it is simple and effective,the taxonomies are difficult to manage because new technologies are introduced rapidly.Therefore,recent studies exploit deep learning to extract pre-defined targets such as problems and solutions.Based on the recent advances in question answering(QA)using deep learning,we adopt a multi-turn QA model to extract problems and solutions from Korean R&D reports.With the previous research,we use the reports directly and analyze the difficulties in handling them using QA style on Information Extraction(IE)for sentence-level benchmark dataset.After investigating the characteristics of Korean R&D,we propose a model to deal with multiple and repeated appearances of targets in the reports.Accordingly,we propose a model that includes an algorithm with two novel modules and a prompt.A newly proposed methodology focuses on reformulating a question without a static template or pre-defined knowledge.We show the effectiveness of the proposed model using a Korean R&D report dataset that we constructed and presented an in-depth analysis of the benefits of the multi-turn QA model.展开更多
Visual question answering(VQA)has attracted more and more attention in computer vision and natural language processing.Scholars are committed to studying how to better integrate image features and text features to ach...Visual question answering(VQA)has attracted more and more attention in computer vision and natural language processing.Scholars are committed to studying how to better integrate image features and text features to achieve better results in VQA tasks.Analysis of all features may cause information redundancy and heavy computational burden.Attention mechanism is a wise way to solve this problem.However,using single attention mechanism may cause incomplete concern of features.This paper improves the attention mechanism method and proposes a hybrid attention mechanism that combines the spatial attention mechanism method and the channel attention mechanism method.In the case that the attention mechanism will cause the loss of the original features,a small portion of image features were added as compensation.For the attention mechanism of text features,a selfattention mechanism was introduced,and the internal structural features of sentences were strengthened to improve the overall model.The results show that attention mechanism and feature compensation add 6.1%accuracy to multimodal low-rank bilinear pooling network.展开更多
The original intention of visual question answering(VQA)models is to infer the answer based on the relevant information of the question text in the visual image,but many VQA models often yield answers that are biased ...The original intention of visual question answering(VQA)models is to infer the answer based on the relevant information of the question text in the visual image,but many VQA models often yield answers that are biased by some prior knowledge,especially the language priors.This paper proposes a mitigation model called language priors mitigation-VQA(LPM-VQA)for the language priors problem in VQA model,which divides language priors into positive and negative language priors.Different network branches are used to capture and process the different priors to achieve the purpose of mitigating language priors.A dynamically-changing language prior feedback objective function is designed with the intermediate results of some modules in the VQA model.The weight of the loss value for each answer is dynamically set according to the strength of its language priors to balance its proportion in the total VQA loss to further mitigate the language priors.This model does not depend on the baseline VQA architectures and can be configured like a plug-in to improve the performance of the model over most existing VQA models.The experimental results show that the proposed model is general and effective,achieving state-of-the-art accuracy in the VQA-CP v2 dataset.展开更多
To improve question answering （QA） performance based on real-world web data sets,a new set of question classes and a general answer re-ranking model are defined.With pre-defined dictionary and grammatical analysis,t...To improve question answering （QA） performance based on real-world web data sets,a new set of question classes and a general answer re-ranking model are defined.With pre-defined dictionary and grammatical analysis,the question classifier draws both semantic and grammatical information into information retrieval and machine learning methods in the form of various training features,including the question word,the main verb of the question,the dependency structure,the position of the main auxiliary verb,the main noun of the question,the top hypernym of the main noun,etc.Then the QA query results are re-ranked by question class information.Experiments show that the questions in real-world web data sets can be accurately classified by the classifier,and the QA results after re-ranking can be obviously improved.It is proved that with both semantic and grammatical information,applications such as QA, built upon real-world web data sets, can be improved,thus showing better performance.展开更多
Aiming at the relation linking task for question answering over knowledge base,especially the multi relation linking task for complex questions,a relation linking approach based on the multi-attention recurrent neural...Aiming at the relation linking task for question answering over knowledge base,especially the multi relation linking task for complex questions,a relation linking approach based on the multi-attention recurrent neural network(RNN)model is proposed,which works for both simple and complex questions.First,the vector representations of questions are learned by the bidirectional long short-term memory(Bi-LSTM)model at the word and character levels,and named entities in questions are labeled by the conditional random field(CRF)model.Candidate entities are generated based on a dictionary,the disambiguation of candidate entities is realized based on predefined rules,and named entities mentioned in questions are linked to entities in knowledge base.Next,questions are classified into simple or complex questions by the machine learning method.Starting from the identified entities,for simple questions,one-hop relations are collected in the knowledge base as candidate relations;for complex questions,two-hop relations are collected as candidates.Finally,the multi-attention Bi-LSTM model is used to encode questions and candidate relations,compare their similarity,and return the candidate relation with the highest similarity as the result of relation linking.It is worth noting that the Bi-LSTM model with one attentions is adopted for simple questions,and the Bi-LSTM model with two attentions is adopted for complex questions.The experimental results show that,based on the effective entity linking method,the Bi-LSTM model with the attention mechanism improves the relation linking effectiveness of both simple and complex questions,which outperforms the existing relation linking methods based on graph algorithm or linguistics understanding.展开更多
A passage retrieval strategy for web-based question answering （QA） systems is proposed in our QA system. It firstly analyzes the question based on semantic patterns to obtain its syntactic and semantic information a...A passage retrieval strategy for web-based question answering （QA） systems is proposed in our QA system. It firstly analyzes the question based on semantic patterns to obtain its syntactic and semantic information and then form initial queries. The queries are used to retrieve documents from the World Wide Web （WWW） using the Google search engine. The queries are then rewritten to form queries for passage retrieval in order to improve the precision. The relations between keywords in the question are employed in our query rewrite method. The experimental result on the question set of the TREC-2003 passage task shows that our system performs well for factoid questions.展开更多
Deep learning models have been shown to have great advantages in answer selection tasks.The existing models,which employ encoder-decoder recurrent neural network(RNN),have been demonstrated to be effective.However,the...Deep learning models have been shown to have great advantages in answer selection tasks.The existing models,which employ encoder-decoder recurrent neural network(RNN),have been demonstrated to be effective.However,the traditional RNN-based models still suffer from limitations such as 1)high-dimensional data representation in natural language processing and 2)biased attentive weights for subsequent words in traditional time series models.In this study,a new answer selection model is proposed based on the Bidirectional Long Short-Term Memory(Bi-LSTM)and attention mechanism.The proposed model is able to generate the more effective question-answer pair representation.Experiments on a question answering dataset that includes information from multiple fields show the great advantages of our proposed model.Specifically,we achieve a maximum improvement of 3.8%over the classical LSTM model in terms of mean average precision.展开更多
In Chinese question answering system, because there is more semantic relation in questions than that in query words, the precision can be improved by expanding query while using natural language questions to retrieve ...In Chinese question answering system, because there is more semantic relation in questions than that in query words, the precision can be improved by expanding query while using natural language questions to retrieve documents. This paper proposes a new approach to query expansion based on semantics and statistics Firstly automatic relevance feedback method is used to generate a candidate expansion word set. Then the expanded query words are selected from the set based on the semantic similarity and seman- tic relevancy between the candidate words and the original words. Experiments show the new approach is effective for Web retrieval and out-performs the conventional expansion approaches.展开更多
With the warming up and continuous development of machine learning,especially deep learning,the research on visual question answering field has made significant progress,with important theoretical research significanc...With the warming up and continuous development of machine learning,especially deep learning,the research on visual question answering field has made significant progress,with important theoretical research significance and practical application value.Therefore,it is necessary to summarize the current research and provide some reference for researchers in this field.This article conducted a detailed and in-depth analysis and summarized of relevant research and typical methods of visual question answering field.First,relevant background knowledge about VQA(Visual Question Answering)was introduced.Secondly,the issues and challenges of visual question answering were discussed,and at the same time,some promising discussion on the particular methodologies was given.Thirdly,the key sub-problems affecting visual question answering were summarized and analyzed.Then,the current commonly used data sets and evaluation indicators were summarized.Next,in view of the popular algorithms and models in VQA research,comparison of the algorithms and models was summarized and listed.Finally,the future development trend and conclusion of visual question answering were prospected.展开更多
The information integration method of semantic web based on agent ontology(SWAO method) was put forward aiming at the problems in current network environment,which integrates,analyzes and processes enormous web inform...The information integration method of semantic web based on agent ontology(SWAO method) was put forward aiming at the problems in current network environment,which integrates,analyzes and processes enormous web information and extracts answers on the basis of semantics. With SWAO method as the clue,the following technologies were studied:the method of concept extraction based on semantic term mining,agent ontology construction method on account of multi-points and the answer extraction in view of semantic inference. Meanwhile,the structural model of the question answering system applying ontology was presented,which adopts OWL language to describe domain knowledge from where QA system infers and extracts answers by Jena inference engine. In the system testing,the precision rate reaches 86%,and the recalling rate is 93%. The experimental results prove that it is feasible to use the method to develop a question answering system,which is valuable for further study in more depth.展开更多
As a new type of knowledge sharing platform,the community question answer website realizes the acquisition and sharing of knowledge,and is loved and sought after by the majority of users.But for multi-answer questions...As a new type of knowledge sharing platform,the community question answer website realizes the acquisition and sharing of knowledge,and is loved and sought after by the majority of users.But for multi-answer questions,answer quality assessment becomes a challenge.The answer selection in CQA(Community Question Answer)was proposed as a challenge task in the SemEval competition,which gave a data set and proposed two subtasks.Task-A is to give a question(including short title and extended description)and its answers,and divide each answer into absolutely relevant(good),potentially relevant(potential)and bad or irrelevant(bad,dialog,non-English,other).Task-B is to give a YES/NO type question(including short title and extended description)and some answers.Based on the answer of the absolute correlation type(good),judge whether the answer to the whole question should be yes,no or uncertain.This paper first preprocesses this data set,and then uses natural language processing technology to perform word segmentation,part-of-speech tagging and named entity recognition on the data set,and then perform feature extraction on the preprocessed data set.Finally,SVM and random forest are used to classify on the basis of feature extraction,and the classification results are analyzed and compared.The experiments in this paper show that SVM and random forest methods have good results on the data set,and exceed the multi-classifier ensemble learning method and hierarchical classification method proposed by the predecessors.展开更多
Commonsense question answering(CQA)requires understanding and reasoning over QA context and related commonsense knowledge,such as a structured Knowledge Graph(KG).Existing studies combine language models and graph neu...Commonsense question answering(CQA)requires understanding and reasoning over QA context and related commonsense knowledge,such as a structured Knowledge Graph(KG).Existing studies combine language models and graph neural networks to model inference.However,traditional knowledge graph are mostly concept-based,ignoring direct path evidence necessary for accurate reasoning.In this paper,we propose MRGNN(Meta-path Reasoning Graph Neural Network),a novel model that comprehensively captures sequential semantic information from concepts and paths.In MRGNN,meta-paths are introduced as direct inference evidence and an original graph neural network is adopted to aggregate features from both concepts and paths simultaneously.We conduct sufficient experiments on the CommonsenceQA and OpenBookQA datasets,showing the effectiveness of MRGNN.Also,we conduct further ablation experiments and explain the reasoning behavior through the case study.展开更多
As a Turing test in multimedia,visual question answering(VQA)aims to answer the textual question with a given image.Recently,the“dynamic”property of neural networks has been explored as one of the most promising way...As a Turing test in multimedia,visual question answering(VQA)aims to answer the textual question with a given image.Recently,the“dynamic”property of neural networks has been explored as one of the most promising ways of improving the adaptability,interpretability,and capacity of the neural network models.Unfortunately,despite the prevalence of dynamic convolutional neural networks,it is relatively less touched and very nontrivial to exploit dynamics in the transformers of the VQA tasks through all the stages in an end-to-end manner.Typically,due to the large computation cost of transformers,researchers are inclined to only apply transformers on the extracted high-level visual features for downstream vision and language tasks.To this end,we introduce a question-guided dynamic layer to the transformer as it can effectively increase the model capacity and require fewer transformer layers for the VQA task.In particular,we name the dynamics in the Transformer as Conditional Multi-Head Self-Attention block(cMHSA).Furthermore,our questionguided cMHSA is compatible with conditional ResNeXt block(cResNeXt).Thus a novel model mixture of conditional gating blocks(McG)is proposed for VQA,which keeps the best of the Transformer,convolutional neural network(CNN),and dynamic networks.The pure conditional gating CNN model and the conditional gating Transformer model can be viewed as special examples of McG.We quantitatively and qualitatively evaluate McG on the CLEVR and VQA-Abstract datasets.Extensive experiments show that McG has achieved the state-of-the-art performance on these benchmark datasets.展开更多
Over the last couple of decades,community question-answering sites(CQAs)have been a topic of much academic interest.Scholars have often leveraged traditional machine learning(ML)and deep learning(DL)to explore the eve...Over the last couple of decades,community question-answering sites(CQAs)have been a topic of much academic interest.Scholars have often leveraged traditional machine learning(ML)and deep learning(DL)to explore the ever-growing volume of content that CQAs engender.To clarify the current state of the CQA literature that has used ML and DL,this paper reports a systematic literature review.The goal is to summarise and synthesise the major themes of CQA research related to(i)questions,(ii)answers and(iii)users.The final review included 133 articles.Dominant research themes include question quality,answer quality,and expert identification.In terms of dataset,some of the most widely studied platforms include Yahoo!Answers,Stack Exchange and Stack Overflow.The scope of most articles was confined to just one platform with few cross-platform investigations.Articles with ML outnumber those with DL.Nonetheless,the use of DL in CQA research is on an upward trajectory.A number of research directions are proposed.展开更多
基金Supported by Sichuan Science and Technology Program(2021YFQ0003,2023YFSY0026,2023YFH0004).
文摘In the field of natural language processing(NLP),there have been various pre-training language models in recent years,with question answering systems gaining significant attention.However,as algorithms,data,and computing power advance,the issue of increasingly larger models and a growing number of parameters has surfaced.Consequently,model training has become more costly and less efficient.To enhance the efficiency and accuracy of the training process while reducing themodel volume,this paper proposes a first-order pruningmodel PAL-BERT based on the ALBERT model according to the characteristics of question-answering(QA)system and language model.Firstly,a first-order network pruning method based on the ALBERT model is designed,and the PAL-BERT model is formed.Then,the parameter optimization strategy of the PAL-BERT model is formulated,and the Mish function was used as an activation function instead of ReLU to improve the performance.Finally,after comparison experiments with traditional deep learning models TextCNN and BiLSTM,it is confirmed that PALBERT is a pruning model compression method that can significantly reduce training time and optimize training efficiency.Compared with traditional models,PAL-BERT significantly improves the NLP task’s performance.
基金supported by Sichuan Science and Technology Program(2023YFSY0026,2023YFH0004).
文摘Recent advancements in natural language processing have given rise to numerous pre-training language models in question-answering systems.However,with the constant evolution of algorithms,data,and computing power,the increasing size and complexity of these models have led to increased training costs and reduced efficiency.This study aims to minimize the inference time of such models while maintaining computational performance.It also proposes a novel Distillation model for PAL-BERT(DPAL-BERT),specifically,employs knowledge distillation,using the PAL-BERT model as the teacher model to train two student models:DPAL-BERT-Bi and DPAL-BERTC.This research enhances the dataset through techniques such as masking,replacement,and n-gram sampling to optimize knowledge transfer.The experimental results showed that the distilled models greatly outperform models trained from scratch.In addition,although the distilled models exhibit a slight decrease in performance compared to PAL-BERT,they significantly reduce inference time to just 0.25%of the original.This demonstrates the effectiveness of the proposed approach in balancing model performance and efficiency.
文摘The weapon and equipment operational requirement analysis(WEORA) is a necessary condition to win a future war,among which the acquisition of knowledge about weapons and equipment is a great challenge. The main challenge is that the existing weapons and equipment data fails to carry out structured knowledge representation, and knowledge navigation based on natural language cannot efficiently support the WEORA. To solve above problem, this research proposes a method based on question answering(QA) of weapons and equipment knowledge graph(WEKG) to construct and navigate the knowledge related to weapons and equipment in the WEORA. This method firstly constructs the WEKG, and builds a neutral network-based QA system over the WEKG by means of semantic parsing for knowledge navigation. Finally, the method is evaluated and a chatbot on the QA system is developed for the WEORA. Our proposed method has good performance in the accuracy and efficiency of searching target knowledge, and can well assist the WEORA.
基金Supported by National Nature Science Foudation of China(61976160,61906137,61976158,62076184,62076182)Shanghai Science and Technology Plan Project(21DZ1204800)。
文摘Background External knowledge representations play an essential role in knowledge-based visual question and answering to better understand complex scenarios in the open world.Recent entity-relationship embedding approaches are deficient in representing some complex relations,resulting in a lack of topic-related knowledge and redundancy in topic-irrelevant information.Methods To this end,we propose MKEAH:Multimodal Knowledge Extraction and Accumulation on Hyperplanes.To ensure that the lengths of the feature vectors projected onto the hyperplane compare equally and to filter out sufficient topic-irrelevant information,two losses are proposed to learn the triplet representations from the complementary views:range loss and orthogonal loss.To interpret the capability of extracting topic-related knowledge,we present the Topic Similarity(TS)between topic and entity-relations.Results Experimental results demonstrate the effectiveness of hyperplane embedding for knowledge representation in knowledge-based visual question answering.Our model outperformed state-of-the-art methods by 2.12%and 3.24%on two challenging knowledge-request datasets:OK-VQA and KRVQA,respectively.Conclusions The obvious advantages of our model in TS show that using hyperplane embedding to represent multimodal knowledge can improve its ability to extract topic-related knowledge.
文摘ExpertRecommendation(ER)aims to identify domain experts with high expertise and willingness to provide answers to questions in Community Question Answering(CQA)web services.How to model questions and users in the heterogeneous content network is critical to this task.Most traditional methods focus on modeling questions and users based on the textual content left in the community while ignoring the structural properties of heterogeneous CQA networks and always suffering from textual data sparsity issues.Recent approaches take advantage of structural proximities between nodes and attempt to fuse the textual content of nodes for modeling.However,they often fail to distinguish the nodes’personalized preferences and only consider the textual content of a part of the nodes in network embedding learning,while ignoring the semantic relevance of nodes.In this paper,we propose a novel framework that jointly considers the structural proximity relations and textual semantic relevance to model users and questions more comprehensively.Specifically,we learn topology-based embeddings through a hierarchical attentive network learning strategy,in which the proximity information and the personalized preference of nodes are encoded and preserved.Meanwhile,we utilize the node’s textual content and the text correlation between adjacent nodes to build the content-based embedding through a meta-context-aware skip-gram model.In addition,the user’s relative answer quality is incorporated to promote the ranking performance.Experimental results show that our proposed framework consistently and significantly outperforms the state-of-the-art baselines on three real-world datasets by taking the deep semantic understanding and structural feature learning together.The performance of the proposed work is analyzed in terms of MRR,P@K,and MAP and is proven to be more advanced than the existing methodologies.
基金supported by the National Research Foundation of Korea(NRF)grant funded by the Korea Government(MSIT)(No.2020R1G1A1100493).
文摘Recently,pre-trained language representation models such as bidirec-tional encoder representations from transformers(BERT)have been performing well in commonsense question answering(CSQA).However,there is a problem that the models do not directly use explicit information of knowledge sources existing outside.To augment this,additional methods such as knowledge-aware graph network(KagNet)and multi-hop graph relation network(MHGRN)have been proposed.In this study,we propose to use the latest pre-trained language model a lite bidirectional encoder representations from transformers(ALBERT)with knowledge graph information extraction technique.We also propose to applying the novel method,schema graph expansion to recent language models.Then,we analyze the effect of applying knowledge graph-based knowledge extraction techniques to recent pre-trained language models and confirm that schema graph expansion is effective in some extent.Furthermore,we show that our proposed model can achieve better performance than existing KagNet and MHGRN models in CommonsenseQA dataset.
基金the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(NRF-2019R1G1A1003312)the Ministry of Education(NRF-2021R1I1A3052815).
文摘Analyzing Research and Development(R&D)trends is important because it can influence future decisions regarding R&D direction.In typical trend analysis,topic or technology taxonomies are employed to compute the popularities of the topics or codes over time.Although it is simple and effective,the taxonomies are difficult to manage because new technologies are introduced rapidly.Therefore,recent studies exploit deep learning to extract pre-defined targets such as problems and solutions.Based on the recent advances in question answering(QA)using deep learning,we adopt a multi-turn QA model to extract problems and solutions from Korean R&D reports.With the previous research,we use the reports directly and analyze the difficulties in handling them using QA style on Information Extraction(IE)for sentence-level benchmark dataset.After investigating the characteristics of Korean R&D,we propose a model to deal with multiple and repeated appearances of targets in the reports.Accordingly,we propose a model that includes an algorithm with two novel modules and a prompt.A newly proposed methodology focuses on reformulating a question without a static template or pre-defined knowledge.We show the effectiveness of the proposed model using a Korean R&D report dataset that we constructed and presented an in-depth analysis of the benefits of the multi-turn QA model.
基金This work was supported by the Sichuan Science and Technology Program(2021YFQ0003).
文摘Visual question answering(VQA)has attracted more and more attention in computer vision and natural language processing.Scholars are committed to studying how to better integrate image features and text features to achieve better results in VQA tasks.Analysis of all features may cause information redundancy and heavy computational burden.Attention mechanism is a wise way to solve this problem.However,using single attention mechanism may cause incomplete concern of features.This paper improves the attention mechanism method and proposes a hybrid attention mechanism that combines the spatial attention mechanism method and the channel attention mechanism method.In the case that the attention mechanism will cause the loss of the original features,a small portion of image features were added as compensation.For the attention mechanism of text features,a selfattention mechanism was introduced,and the internal structural features of sentences were strengthened to improve the overall model.The results show that attention mechanism and feature compensation add 6.1%accuracy to multimodal low-rank bilinear pooling network.
文摘The original intention of visual question answering(VQA)models is to infer the answer based on the relevant information of the question text in the visual image,but many VQA models often yield answers that are biased by some prior knowledge,especially the language priors.This paper proposes a mitigation model called language priors mitigation-VQA(LPM-VQA)for the language priors problem in VQA model,which divides language priors into positive and negative language priors.Different network branches are used to capture and process the different priors to achieve the purpose of mitigating language priors.A dynamically-changing language prior feedback objective function is designed with the intermediate results of some modules in the VQA model.The weight of the loss value for each answer is dynamically set according to the strength of its language priors to balance its proportion in the total VQA loss to further mitigate the language priors.This model does not depend on the baseline VQA architectures and can be configured like a plug-in to improve the performance of the model over most existing VQA models.The experimental results show that the proposed model is general and effective,achieving state-of-the-art accuracy in the VQA-CP v2 dataset.
基金Microsoft Research Asia Internet Services in Academic Research Fund(No.FY07-RES-OPP-116)the Science and Technology Development Program of Tianjin(No.06YFGZGX05900)
文摘To improve question answering （QA） performance based on real-world web data sets,a new set of question classes and a general answer re-ranking model are defined.With pre-defined dictionary and grammatical analysis,the question classifier draws both semantic and grammatical information into information retrieval and machine learning methods in the form of various training features,including the question word,the main verb of the question,the dependency structure,the position of the main auxiliary verb,the main noun of the question,the top hypernym of the main noun,etc.Then the QA query results are re-ranked by question class information.Experiments show that the questions in real-world web data sets can be accurately classified by the classifier,and the QA results after re-ranking can be obviously improved.It is proved that with both semantic and grammatical information,applications such as QA, built upon real-world web data sets, can be improved,thus showing better performance.
基金The National Natural Science Foundation of China(No.61502095).
文摘Aiming at the relation linking task for question answering over knowledge base,especially the multi relation linking task for complex questions,a relation linking approach based on the multi-attention recurrent neural network(RNN)model is proposed,which works for both simple and complex questions.First,the vector representations of questions are learned by the bidirectional long short-term memory(Bi-LSTM)model at the word and character levels,and named entities in questions are labeled by the conditional random field(CRF)model.Candidate entities are generated based on a dictionary,the disambiguation of candidate entities is realized based on predefined rules,and named entities mentioned in questions are linked to entities in knowledge base.Next,questions are classified into simple or complex questions by the machine learning method.Starting from the identified entities,for simple questions,one-hop relations are collected in the knowledge base as candidate relations;for complex questions,two-hop relations are collected as candidates.Finally,the multi-attention Bi-LSTM model is used to encode questions and candidate relations,compare their similarity,and return the candidate relation with the highest similarity as the result of relation linking.It is worth noting that the Bi-LSTM model with one attentions is adopted for simple questions,and the Bi-LSTM model with two attentions is adopted for complex questions.The experimental results show that,based on the effective entity linking method,the Bi-LSTM model with the attention mechanism improves the relation linking effectiveness of both simple and complex questions,which outperforms the existing relation linking methods based on graph algorithm or linguistics understanding.
基金Supported by the National Basic Research Program of China (2003CB317002)the Grant from City University of Hong Kong (7002137)
文摘A passage retrieval strategy for web-based question answering （QA） systems is proposed in our QA system. It firstly analyzes the question based on semantic patterns to obtain its syntactic and semantic information and then form initial queries. The queries are used to retrieve documents from the World Wide Web （WWW） using the Google search engine. The queries are then rewritten to form queries for passage retrieval in order to improve the precision. The relations between keywords in the question are employed in our query rewrite method. The experimental result on the question set of the TREC-2003 passage task shows that our system performs well for factoid questions.
基金This work was supported in part by the National Natural Science Foundation of China under Grant 61572326,and Grant 61802258the Natural Science Foundation of Shanghai under Grant 18ZR1428300the Shanghai Committee of Science and Technology under Grant 17070502800 and Grant 16JC1403000.
文摘Deep learning models have been shown to have great advantages in answer selection tasks.The existing models,which employ encoder-decoder recurrent neural network(RNN),have been demonstrated to be effective.However,the traditional RNN-based models still suffer from limitations such as 1)high-dimensional data representation in natural language processing and 2)biased attentive weights for subsequent words in traditional time series models.In this study,a new answer selection model is proposed based on the Bidirectional Long Short-Term Memory(Bi-LSTM)and attention mechanism.The proposed model is able to generate the more effective question-answer pair representation.Experiments on a question answering dataset that includes information from multiple fields show the great advantages of our proposed model.Specifically,we achieve a maximum improvement of 3.8%over the classical LSTM model in terms of mean average precision.
基金the Specialized Research Program Fundthe Doctoral Program of Higher Education of China (20050007023)the Natural Science Foundation of Shandong Province(Y2004G04)
文摘In Chinese question answering system, because there is more semantic relation in questions than that in query words, the precision can be improved by expanding query while using natural language questions to retrieve documents. This paper proposes a new approach to query expansion based on semantics and statistics Firstly automatic relevance feedback method is used to generate a candidate expansion word set. Then the expanded query words are selected from the set based on the semantic similarity and seman- tic relevancy between the candidate words and the original words. Experiments show the new approach is effective for Web retrieval and out-performs the conventional expansion approaches.
基金Project(61702063)supported by the National Natural Science Foundation of China。
文摘With the warming up and continuous development of machine learning,especially deep learning,the research on visual question answering field has made significant progress,with important theoretical research significance and practical application value.Therefore,it is necessary to summarize the current research and provide some reference for researchers in this field.This article conducted a detailed and in-depth analysis and summarized of relevant research and typical methods of visual question answering field.First,relevant background knowledge about VQA(Visual Question Answering)was introduced.Secondly,the issues and challenges of visual question answering were discussed,and at the same time,some promising discussion on the particular methodologies was given.Thirdly,the key sub-problems affecting visual question answering were summarized and analyzed.Then,the current commonly used data sets and evaluation indicators were summarized.Next,in view of the popular algorithms and models in VQA research,comparison of the algorithms and models was summarized and listed.Finally,the future development trend and conclusion of visual question answering were prospected.
基金Projects(60773462, 60672171) supported by the National Natural Science Foundation of ChinaProjects(2009AA12143, 2009AA012136) supported by the National High-Tech Research and Development Program of ChinaProject(20080430250) supported by the Foundation of Post-Doctor in China
文摘The information integration method of semantic web based on agent ontology(SWAO method) was put forward aiming at the problems in current network environment,which integrates,analyzes and processes enormous web information and extracts answers on the basis of semantics. With SWAO method as the clue,the following technologies were studied:the method of concept extraction based on semantic term mining,agent ontology construction method on account of multi-points and the answer extraction in view of semantic inference. Meanwhile,the structural model of the question answering system applying ontology was presented,which adopts OWL language to describe domain knowledge from where QA system infers and extracts answers by Jena inference engine. In the system testing,the precision rate reaches 86%,and the recalling rate is 93%. The experimental results prove that it is feasible to use the method to develop a question answering system,which is valuable for further study in more depth.
文摘As a new type of knowledge sharing platform,the community question answer website realizes the acquisition and sharing of knowledge,and is loved and sought after by the majority of users.But for multi-answer questions,answer quality assessment becomes a challenge.The answer selection in CQA(Community Question Answer)was proposed as a challenge task in the SemEval competition,which gave a data set and proposed two subtasks.Task-A is to give a question(including short title and extended description)and its answers,and divide each answer into absolutely relevant(good),potentially relevant(potential)and bad or irrelevant(bad,dialog,non-English,other).Task-B is to give a YES/NO type question(including short title and extended description)and some answers.Based on the answer of the absolute correlation type(good),judge whether the answer to the whole question should be yes,no or uncertain.This paper first preprocesses this data set,and then uses natural language processing technology to perform word segmentation,part-of-speech tagging and named entity recognition on the data set,and then perform feature extraction on the preprocessed data set.Finally,SVM and random forest are used to classify on the basis of feature extraction,and the classification results are analyzed and compared.The experiments in this paper show that SVM and random forest methods have good results on the data set,and exceed the multi-classifier ensemble learning method and hierarchical classification method proposed by the predecessors.
基金supported by the Key Research and Development Program of Hubei Province(2020BAB017)the Scientific Research Center Program of National Language Commission(ZDI135-135)the Fundamental Research Funds for the Central Universities(KJ02502022-0155,CCNU22XJ037).
文摘Commonsense question answering(CQA)requires understanding and reasoning over QA context and related commonsense knowledge,such as a structured Knowledge Graph(KG).Existing studies combine language models and graph neural networks to model inference.However,traditional knowledge graph are mostly concept-based,ignoring direct path evidence necessary for accurate reasoning.In this paper,we propose MRGNN(Meta-path Reasoning Graph Neural Network),a novel model that comprehensively captures sequential semantic information from concepts and paths.In MRGNN,meta-paths are introduced as direct inference evidence and an original graph neural network is adopted to aggregate features from both concepts and paths simultaneously.We conduct sufficient experiments on the CommonsenceQA and OpenBookQA datasets,showing the effectiveness of MRGNN.Also,we conduct further ablation experiments and explain the reasoning behavior through the case study.
基金supported in part by the National Natural Science Foundation of China under Grant No.62176061the Science and Technology Commission of Shanghai Municipality under Grant No.22511105000.
文摘As a Turing test in multimedia,visual question answering(VQA)aims to answer the textual question with a given image.Recently,the“dynamic”property of neural networks has been explored as one of the most promising ways of improving the adaptability,interpretability,and capacity of the neural network models.Unfortunately,despite the prevalence of dynamic convolutional neural networks,it is relatively less touched and very nontrivial to exploit dynamics in the transformers of the VQA tasks through all the stages in an end-to-end manner.Typically,due to the large computation cost of transformers,researchers are inclined to only apply transformers on the extracted high-level visual features for downstream vision and language tasks.To this end,we introduce a question-guided dynamic layer to the transformer as it can effectively increase the model capacity and require fewer transformer layers for the VQA task.In particular,we name the dynamics in the Transformer as Conditional Multi-Head Self-Attention block(cMHSA).Furthermore,our questionguided cMHSA is compatible with conditional ResNeXt block(cResNeXt).Thus a novel model mixture of conditional gating blocks(McG)is proposed for VQA,which keeps the best of the Transformer,convolutional neural network(CNN),and dynamic networks.The pure conditional gating CNN model and the conditional gating Transformer model can be viewed as special examples of McG.We quantitatively and qualitatively evaluate McG on the CLEVR and VQA-Abstract datasets.Extensive experiments show that McG has achieved the state-of-the-art performance on these benchmark datasets.
文摘Over the last couple of decades,community question-answering sites(CQAs)have been a topic of much academic interest.Scholars have often leveraged traditional machine learning(ML)and deep learning(DL)to explore the ever-growing volume of content that CQAs engender.To clarify the current state of the CQA literature that has used ML and DL,this paper reports a systematic literature review.The goal is to summarise and synthesise the major themes of CQA research related to(i)questions,(ii)answers and(iii)users.The final review included 133 articles.Dominant research themes include question quality,answer quality,and expert identification.In terms of dataset,some of the most widely studied platforms include Yahoo!Answers,Stack Exchange and Stack Overflow.The scope of most articles was confined to just one platform with few cross-platform investigations.Articles with ML outnumber those with DL.Nonetheless,the use of DL in CQA research is on an upward trajectory.A number of research directions are proposed.
