检索结果-内蒙古大学图书馆

12th workshop on graph-based methods for natural language processing, Textgraphs 2018 - in conjunction with the 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human, NAACL HLT 2018

ISBN: (纸本)9781948087254

The proceedings contain 8 papers. The topics discussed include: embedding text in hyperbolic spaces;scientific discovery as link prediction in influence and citation graphs;efficient generation and processing of word co-occurrence networks using corpus2graph;multi-hop inference for sentence-level Textgraphs: how challenging is meaningfully combining information for science question answering?;multi-sentence compression with word vertex-labeled graphs and integer linear programming;large-scale spectral clustering using diffusion coordinates on landmark-based bipartite graphs;efficient graph-based word sense induction by distributional inclusion vector embeddings;fusing document, collection and label graph-based representations with word embeddings for text classification;and embedding text in hyperbolic spaces.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Optimizing Symptom and Disease Representation with Structural and Textual Information Integration

Optimizing Symptom and Disease Representation with Structura...

引用

Multi-Agent Systems for Collaborative Intelligence (ICMSCI), International Conference on

作者： M. Ramya Meenakshi B K Kanagalakshmi M Department Of CSE St. Joseph's Institute of Technology Chennai-119

ISBN: (数字)9798331509828

ISBN: (纸本)9798331509835

Medical websites and healthcare records now contain an enormous variety of health-related data from online and offline medical check-ins and consultations. However, because there are few symptom-disease connections, it is difficult to leverage user-generated material to help patients get around online and receive medical attention offline. The casual and lighthearted ways that symptoms are expressed in the data are what lead to the severe sparsity. To improve therapeutic interventions and diagnostic systems, accurate portrayal of diseases and symptoms is essential. Medical websites and healthcare records now contain an enormous variety of health-related data from online and offline medical check-ins and consultations. However, because there are few symptom-disease connections, it is difficult to leverage content created by users to help patients get around online and receive healthcare offline. The casual and light hearted ways that symptoms are expressed in the data are what lead to the severe sparsity. To improve therapies and diagnostic systems, accurate portrayal of diseases and symptoms is essential. Our method involves a multi-phase approach: first, extracting and encoding structured information from EHRs using advanced techniques such as graph-based neural networks; second, applying natural language processing (NLP) techniques to capture and encode relevant information from clinical texts; and finally, fusing these two modalities to create an enriched representation of medical entities. We evaluate our approach on a dataset of medical records and clinical notes, demonstrating improvements in predictive accuracy and interpretability compared to traditional single-modality methods. The results highlight the effectiveness of combining structural and textual data in advancing the state-of-the-art in medical representation learning. This fusion approach not only enhances the precision of symptom and disease modeling but also provides deeper insights into complex m

关键词： Accuracy Smart healthcare Medical services Prediction algorithms Throughput natural language processing Sensors Data mining Medical diagnostic imaging Diseases

来源：评论

学校读者我要写书评

暂无评论

Generative non-autoregressive unsupervised keyphrase extraction with neural topic modeling

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2023年第1期120卷

作者： Zhu, Xun Lou, Yinxia Zhao, Jing Gao, Wang Deng, Hongtao Jianghan Univ Sch Artificial Intelligence Wuhan Peoples R China Jianghan Univ Engn Res Ctr Intelligent Decis & Informat Proc Wuhan Peoples R China

Unsupervised keyphrase extraction (UKE) aims to detect out a set of keyphrases of a document without using any annotation signal for training the UKE model. Existing UKE models however either fail to solve the issue of low keyphrase relevance, or suffer from the problem of unsatisfactory keyphrase coverage. Besides, the current strong-performing neural based UKE methods unfortunately come with the cost of lower decoding efficiency. In this paper we propose a novel framework for UKE. We model the task as a fully generative process with an encoder-decoder framework, which unifies the prediction of both the present and absent keyphrases in an end-to-end manner. We accelerate the decoding by installing a non-autoregressive decoder, which yields all keyphrase extraction in parallel fast. To enhance the keyphrase relevance, we investigate a neural topic module that unsupervisingly learns the latent topic information, aiding the semantic understanding of the input document. To strengthen the keyphrase coverage, we present an anchor-region scheme for the extraction of present keyphrases, which first locates the anchor points of present phrases and then determines the final boundaries of keyphrases. The initial anchor phrases are organized into an anchor-aware graph, which is modeled by our proposed topic-guided graph attention network which captures the document contexts at global level. Over six benchmark datasets of UKE our system shows better experimental results than existing strong-performing baselines, becoming the new state-of-the-art UKE model. Further in-depth analyses show that our model can effectively resolve both the relevance and coverage of keyphrase extraction, meanwhile being faster on decoding.

关键词： natural language processing Keyphrase extraction Unsupervised learning Topic modeling graph attention network

来源：评论

学校读者我要写书评

暂无评论

Self-Pro: A Self-Prompt and Tuning Framework for graph Neural Networks

arXiv

引用

arXiv 2023年

作者： Gong, Chenghua Li, Xiang Yu, Jianxiang Cheng, Yao Tan, Jiaqi Yu, Chengcheng School of Data Science and Engineering East China Normal University China School of Computer and Information Engineering Shanghai Polytechnic University China

graphs have become an important modeling tool for web applications, and graph Neural Networks (GNNs) have achieved great success in graph representation learning. However, the performance of traditional GNNs heavily relies on a large amount of supervision. Recently, "pre-train, fine-tune" has become the paradigm to address the issues of label dependency and poor generalization. However, the pretraining strategies vary for graphs with homophily and heterophily, and the objectives for various downstream tasks also differ. This leads to a gap between pretexts and downstream tasks, resulting in "negative transfer" and poor performance. Inspired by prompt learning in natural language processing (NLP), many studies turn to bridge the gap and fully leverage the pre-trained model. However, existing methods for graph prompting are tailored to homophily, neglecting inherent heterophily on graphs. Meanwhile, most of them rely on the randomly initialized prompts, which negatively impact on the stability. Therefore, we propose Self-Prompt, a prompting framework for graphs based on the model and data itself. We first introduce asymmetric graph contrastive learning for pretext to address heterophily and align the objectives of pretext and downstream tasks. Then we reuse the component from pre-training phase as the self adapter and introduce self-prompts based on graph itself for task adaptation. Finally, we conduct extensive experiments on 11 benchmark datasets to demonstrate its superiority. We provide our codes at https://***/gongchenghua/Self-Pro. Copyright © 2023, The Authors. All rights reserved.

关键词： graph neural networks

来源：评论

学校读者我要写书评

暂无评论

KERE: A General Reasoning Engine for Complex Problem Solving

KERE: A General Reasoning Engine for Complex Problem Solving

引用

Electronic Communication and Artificial Intelligence (IWECAI), International workshop on

作者： Haowei Cao Yin Chen School of Computer Science South China Normal University Guangzhou China School of Artificial Intelligence South China Normal University Guangzhou China

ISBN: (数字)9798350386943

ISBN: (纸本)9798350386950

Reasoning ability is a critical aspect of human intelligence and a focal point in artificial intelligence research. Currently, both data-driven and knowledge-driven methods face challenges in reasoning tasks. This paper proposes a novel approach that integrates data-driven methods into knowledge-driven reasoning systems. The research encompasses the following aspects: first, we introduce a knowledge-based reasoning system comprising knowledge preprocessing, reasoning engine, and knowledge post-processing modules. The reasoning engine, knowledge equation reasoning engine (KERE), driven by knowledge, is the core component, while the remaining modules serve as tools connecting the reasoning engine and natural language, driven by data. KERE is based on assertion logic theory and knowledge equation. We introduce the basic concepts of assertion logic and provide a formal definition of knowledge equation. Next, we explore the implementation of the KERE. KERE is designed using a production system-like framework, employing knowledge-action pairs and an improved pattern matching algorithm, Rete-KE. Detailed implementation and pseudocode are provided. Finally, we elucidate KERE's operational mechanisms further through case studies in mathematical problem-solving and financial reasoning tasks. Evaluation on the open-source dataset Conic10K demonstrates KERE's significant advantages over purely data-driven methods in mathematical reasoning tasks. Additionally, we discuss initial explorations of applying KERE in network automation operations, demonstrating its potential in various domain-specific reasoning tasks.

关键词： Automation Large language models Knowledge acquisition Writing Cognition Problem-solving Logic

来源：评论

学校读者我要写书评

暂无评论

Joint biomedical entity and relation extraction with knowledge-enhanced collective inference 59

Joint biomedical entity and relation extraction with knowled...

引用

Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on natural language processing, ACL-IJCNLP 2021

作者： Lai, Tuan Ji, Heng Zhai, ChengXiang Tran, Quan Hung University of Illinois at Urbana-Champaign United States Adobe Research

ISBN: (纸本)9781954085527

Compared to the general news domain, information extraction (IE) from biomedical text requires much broader domain knowledge. However, many previous IE methods do not utilize any external knowledge during inference. Due to the exponential growth of biomedical publications, models that do not go beyond their fixed set of parameters will likely fall behind. Inspired by how humans look up relevant information to comprehend a scientific text, we present a novel framework that utilizes external knowledge for joint entity and relation extraction named KECI (Knowledge-Enhanced Collective Inference). Given an input text, KECI first constructs an initial span graph representing its initial understanding of the text. It then uses an entity linker to form a knowledge graph containing relevant background knowledge for the the entity mentions in the text. To make the final predictions, KECI fuses the initial span graph and the knowledge graph into a more refined graph using an attention mechanism. KECI takes a collective approach to link mention spans to entities by integrating global relational information into local representations using graph convolutional networks. Our experimental results show that the framework is highly effective, achieving new state-of-the-art results in two different benchmark datasets: BioRelEx (binding interaction detection) and ADE (adverse drug event extraction). For example, KECI achieves absolute improvements of 4.59% and 4.91% in F1 scores over the state-of-the-art on the BioRelEx entity and relation extraction tasks. © 2021 Association for Computational Linguistics

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Improving graph-based Text Representations with Character and Word Level N-grams

arXiv

引用

arXiv 2022年

作者： Li, Wenzhe Aletras, Nikolaos Computer Science Department University of Sheffield United Kingdom

graph-based text representation focuses on how text documents are represented as graphs for exploiting dependency information between tokens and documents within a corpus. Despite the increasing interest in graph representation learning, there is limited research in exploring new ways for graph-based text representation, which is important in downstream natural language processing tasks. In this paper, we first propose a new heterogeneous word-character text graph that combines word and character n-gram nodes together with document nodes, allowing us to better learn dependencies among these entities. Additionally, we propose two new graph-based neural models, WCTextGCN and WCTextGAT, for modeling our proposed text graph. Extensive experiments in text classification and automatic text summarization benchmarks demonstrate that our proposed models consistently outperform competitive baselines and state-of-the-art graph-based models. © 2022, CC BY.

关键词： graphic methods

来源：评论

学校读者我要写书评

暂无评论

Drug repositioning based on network-specific core genes identifies potential drugs for the treatment of autism spectrum disorder in children

引用

COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL 2021年 19卷 3908-3921页

作者： Gao, Huan Ni, Yuan Mo, Xueying Li, Dantong Teng, Shan Huang, Qingsheng Huang, Shuai Liu, Guangjian Zhang, Sheng Tang, Yaping Lu, Long Liang, Huiying Guangdong Prov Peoples Hosp Guangdong Acad Med Sci Clin Data Ctr Guangzhou 510080 Guangdong Peoples R China Wuhan Univ Sch Informat Management Wuhan 430072 Hubei Peoples R China Ping An Technol 20 Keji South 12 Rd Shen Zhen 518063 Guangdong Peoples R China Guangzhou Med Univ Guangzhou Women & Childrens Med Ctr Inst Pediat 9 Jinsui Rd Guangzhou 510623 Guangdong Peoples R China Southern Med Univ Sch Publ Hlth Dept Psychol Guangzhou 510515 Guangdong Peoples R China

Identification of exact causative genes is important for in silico drug repositioning based on drug-gene disease relationships. However, the complex polygenic etiology of the autism spectrum disorder (ASD) is a challenge in the identification of etiological genes. The network-based core gene identification method can effectively use the interactions between genes and accurately identify the pathogenic genes of ASD. We developed a novel network-based drug repositioning framework that contains three steps: network-specific core gene (NCG) identification, potential therapeutic drug repositioning, and candidate drug validation. first, through the analysis of transcriptome data for 178 brain tissues, gene network analysis identified 365 NCGs in 18 coexpression modules that were significantly correlated with ASD. Second, we evaluated two proposed drug repositioning methods. In one novel approach (dtGSEA), we used the NCGs to probe drug-gene interaction data and identified 35 candidate drugs. In another approach, we compared NCG expression patterns with drug-induced transcriptome data from the Connectivity Map database and found 46 candidate drugs. Third, we validated the candidate drugs using an in-house mental diseases and compounds knowledge graph (MCKG) that contained 7509 compounds, 505 mental diseases, and 123,890 edges. We found a total of 42 candidate drugs that were associated with mental illness, among which 10 drugs (baclofen, sulpiride, estradiol, entinostat, everolimus, fluvoxamine, curcumin, calcitriol, metronidazole, and zinc) were postulated to be associated with ASD. This study proposes a powerful network-based drug repositioning framework and also provides candidate drugs as well as potential drug targets for the subsequent development of ASD therapeutic drugs. (C) 2021 The Authors. Published by Elsevier B.V. on behalf of Research Network of Computational and Structural Biotechnology.

关键词： Autism spectrum disorder Coexpression network Drug repositioning Knowledge graph natural language processing

来源：评论

学校读者我要写书评

暂无评论

Space Target Pose Estimation based on High-Level Geometric Correspondences for Single-Photon LiDAR

Space Target Pose Estimation Based on High-Level Geometric C...

引用

IEEE International workshop on Imaging Systems and Techniques (IST)

作者： Xinyu Wang Duan Li Linjie Lyu Lijun Xu School of Instrumentation and Optoelectronic Engineering Beihang University Beijing China

ISBN: (数字)9798350378214

ISBN: (纸本)9798350378221

Single-photon LiDAR is considered to be one of the most promising techniques for space target pose estimation due to its low power consumption, small size, and lightweight. And it is widely used in space target rendezvous, docking, and landing. However, such processing needs to address plenty of high-noise photon point clouds with limited resources and be completed in a very short time. To address these challenges, a global optimization method of pose estimation based on high-level geometric correspondences is proposed in this work. first, a K-Nearest Neighbors undirected graph is designed for describing the geometrical consistency of correspondences and a degree matrix is calculated to remove the false correspondences. In addition, a robust error metric based on Cauchy's function is introduced, which can effectively speed up the convergence of the optimization and iteratively update the parameters of the transformation matrix. The process of scanning a satellite with the single-photon LiDAR to obtain photon point clouds was simulated using the Monte Carlo method. Simulation results show that the proposed method can accurately estimate the pose of the satellite with a noise rate of 5%, 10%, and 15% in a time less than 1 second, revealing that it is of an obvious advantage in both efficiency and robustness when dealing with large-scale and noisy photon point clouds.

关键词： Point cloud compression Laser radar Satellites Power demand Simulation Pose estimation Noise Optimization methods Robustness Photonics

来源：评论

学校读者我要写书评

暂无评论

Multi-Stream graph Convolutional Networks for Text Classification via Representative-Word Document Mining

引用

INTERNATIONAL JOURNAL OF COMPUTATIONAL INTELLIGENCE AND APPLICATIONS 2022年第4期21卷 2250028-2250028页

作者： Li, Meng Chen, Shenyu Yang, Weifeng Wang, Qianying Hebei Univ Econ & Business Coll Math & Stat Hebei Peoples R China Alibaba Grp Hangzhou Zhejiang Peoples R China

Recently, graph convolutional networks (GCNs) for text classification have received considerable attention in natural language processing. However, most current methods just use original documents and words in the corpus to construct the topology of graph which may lose some effective information. In this paper, we propose a Multi-Stream graph Convolutional Network (MS-GCN) for text classification via Representative-Word Document (RWD) mining, which is implemented in PyTorch. In the proposed method, we first introduce temporary labels and mine the RWDs which are treated as additional documents in the corpus. Then, we build a heterogeneous graph based on relations among a Group of RWDs (GRWDs), words and original documents. Furthermore, we construct the MS-GCN based on multiple heterogeneous graphs according to different GRWDs. Finally, we optimize our MS-GCN model through updated mechanism of GRWDs. We evaluate the proposed approach on six text classification datasets, 20NG, R8, R52, Ohsumed, MR and Pheme. Extensive experiments on these datasets show that our proposed approach outperforms state-of-the-art methods for text classification.

关键词： Text classification neural network heterogeneous graph corpus graph convolution

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：