检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Fu, Lingyue Chai, Huacan Du, Kounianhua Zhang, Weiming Luo, Shuang Lin, Jianghao Fang, Yuchen Rui, Renting Guan, Hao Liu, Jianxing Qi, Siyuan Fan, Longteng Lei, Jiayi Liu, Yifan Wang, Jingkuan Zhang, Kangning Zhang, Weinan Yu, Yong Apex Data & Knowledge Management Lab Shanghai Jiao Tong University China

With the emergence of Large Language Models (LLMs), there has been a significant improvement in the programming capabilities of models, attracting growing attention from researchers. Evaluating the programming capabilities of LLMs is crucial as it reflects the multifaceted abilities of LLMs, and it has numerous downstream applications. In this paper, we propose Codeapex, a bilingual benchmark dataset focusing on the programming comprehension, code generation, and code correction abilities of LLMs. Programming comprehension task tests LLMs on multiple-choice exam questions covering conceptual understanding, commonsense reasoning, and multi-hop reasoning. The code generation task evaluates LLMs through completing C++ functions based on provided descriptions and prototypes. The code correction task asks LLMs to fix real-world erroneous code segments with different error messages. We evaluate 12 widely used LLMs, including both general-purpose and specialized models. GPT-4 exhibits the best programming capabilities, achieving approximate accuracy of 69%, 54%, and 66% on the three tasks, respectively. Compared to human performance, there is still significant room for improvement in LLM programming. We hope that Codeapex can serve as a reference for evaluating the coding capabilities of LLMs, further promoting their development and growth. Copyright © 2023, The Authors. All rights reserved.

关键词： C++ (programming language)

来源：评论

学校读者我要写书评

暂无评论

QA4IE: A Question Answering Based System for Document-Level General Information Extraction

引用

IEEE ACCESS 2020年 8卷 29677-29689页

作者： Qiu, Lin Ru, Dongyu Long, Quanyu Zhang, Weinan Yu, Yong Shanghai Jiao Tong Univ APEX Data & Knowledge Management Lab Shanghai 200240 Peoples R China

Information Extraction (IE) is the task of distilling structured information from unstructured texts by identifying references to named entities as well as relationships between such entities. Existing IE solutions, including Relation Extraction and Open IE, can hardly take cross-sentence information like coreferences into account and are severely restricted by limited relation types as well as informal relation specifications (e.g., free-text based relation triples). In order to overcome the weaknesses, we propose a novel IE framework named QA4IE, which leverages the flexible question answering approaches to produce high-quality relation triples across sentences. Based on this framework, we develop a real-time IE system, which can perform general IE throughout the entire document. For training and evaluating our system, we build a large-scale IE benchmark using distant supervision under human evaluation. We deploy both component analyses and pipeline experiments to evaluate our system. The results show that our system can generalize on unseen entities and relations, as well as achieve significant improvements over existing IE systems.

关键词： knowledge acquisition machine learning natural language processing neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep Recurrent Survival Analysis 33

Deep Recurrent Survival Analysis

引用

33rd AAAI Conference on Artificial Intelligence / 31st Innovative Applications of Artificial Intelligence Conference / 9th AAAI Symposium on Educational Advances in Artificial Intelligence

作者： Ren, Kan Qin, Jiarui Zheng, Lei Yang, Zhengyu Zhang, Weinan Qiu, Lin Yu, Yong Shanghai Jiao Tong Univ APEX Data & Knowledge Management Lab Shanghai Peoples R China

ISBN: (纸本)9781577358091

Survival analysis is a hotspot in statistical research for modeling time-to-event information with data censorship handling, which has been widely used in many applications such as clinical research, information system and other fields with survivorship bias. Many works have been proposed for survival analysis ranging from traditional statistic methods to machine learning models. However, the existing methodologies either utilize counting-based statistics on the segmented data, or have a pre-assumption on the event probability distribution w.r.t. time. Moreover, few works consider sequential patterns within the feature space. In this paper, we propose a Deep Recurrent Survival Analysis model which combines deep learning for conditional probability prediction at fine-grained level of the data, and survival analysis for tackling the censorship. By capturing the time dependency through modeling the conditional probability of the event for each sample, our method predicts the likelihood of the true event occurrence and estimates the survival rate over time, i.e., the probability of the non-occurrence of the event, for the censored data. Meanwhile, without assuming any specific form of the event probability distribution, our model shows great advantages over the previous works on fitting various sophisticated data distributions. In the experiments on the three real-world tasks from different fields, our model significantly outperforms the state-of-the-art solutions under various metrics.

关键词： Probability distributions

来源：评论

学校读者我要写书评

暂无评论

A Bootstrapping Framework With Interactive Information Modeling for Network Alignment

引用

IEEE ACCESS 2018年 6卷 13685-13696页

作者： Cao, Xuezhi Zhang, Weinan Yu, Yong Shanghai Jiao Tong Univ APEX Data & Knowledge Management Lab Shanghai 200240 Peoples R China

Network alignment is an emerging research topic with great utility value. Its purpose is to recover the hidden alignment between different networks for subsequent cross-network research. Its applications benefit numerous domains including online social networks, biological researches, and so on. Hence, recently, it draws plenty of research attention. Existing attempts indicate that it is a promising direction. However, due to its recency, there are still several limitations. Existing literature mostly focuses on network alignment in specific scenarios and employs heuristically designed functions;thus, it does not provide general solutions. Besides, they only capture static information but not interactive information that is based not only on the isolated networks but also the alignment itself. In this paper, we propose a bootstrapping framework for network alignment (BoNA). Specifically, the framework starts from an imperfect seed alignment and then iteratively refines it by capturing interactive information until convergence. Within each iteration, we calculate the likelihoods of pairwise alignment by using supervised learning techniques;hence, heuristic functions are no longer required. Furthermore, we extend the framework by starting from multiple seed alignments to combine their advantages. We conduct experiments in both biology and online social networks to demonstrate the generality of our framework. Results indicate that BoNA outperforms existing aligners significantly in all evaluated scenarios in terms of both node correctness and topology conservation.

关键词： Bootstrapping network alignment online social network protein-protein interaction network

来源：评论

学校读者我要写书评

暂无评论

Neural Link Prediction Over Aligned Networks 32

Neural Link Prediction Over Aligned Networks

引用

32nd AAAI Conference on Artificial Intelligence / 30th Innovative Applications of Artificial Intelligence Conference / 8th AAAI Symposium on Educational Advances in Artificial Intelligence

作者： Cao, Xuezhi Chen, Haokun Wang, Xuejian Zhang, Weinan Yu, Yong Shanghai Jiao Tong Univ APEX Data & Knowledge Management Lab Shanghai Peoples R China

ISBN: (纸本)9781577358008

Link prediction is a fundamental problem with a wide range of applications in various domains, which predicts the links that are not yet observed or the links that may appear in the future. Most existing works in this field only focus on modeling a single network, while real-world networks are actually aligned with each other. Network alignments contain valuable additional information for understanding the networks, and provide a new direction for addressing data insufficiency and alleviating cold start problem. However, there are rare works leveraging network alignments for better link prediction. Besides, neural network is widely employed in various domains while its capability of capturing high-level patterns and correlations for link prediction problem has not been adequately researched yet. Hence, in this paper we target at link prediction over aligned networks using neural networks. The major challenge is the heterogeneousness of the considered networks, as the networks may have different characteristics, link purposes, etc. To overcome this, we propose a novel multi-neural-network framework MNN, where we have one individual neural network for each heterogeneous target or feature while the vertex representations are shared. We further discuss training methods for the multi-neural-network framework. Extensive experiments demonstrate that MNN outperforms the state-of-the-art methods and achieves 3% to 5% relative improvement of AUC score across different settings, particularly over 8% for cold start scenarios.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

label-Aware double transfer learning for cross-specialty medical named entity recognition

Label-Aware double transfer learning for cross-specialty med...

引用

2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2018

作者： Wang, Zhenghui Qu, Yanru Chen, Liheng Shen, Jian Zhang, Weinan Zhang, Shaodian Gao, Yimei Gu, Gen Chen, Ken Yu, Yong APEX Data and Knowledge Management Lab Shanghai Jiao Tong University China Synyi LLC.

ISBN: (纸本)9781948087278

We study the problem of named entity recognition (NER) from electronic medical records, which is one of the most fundamental and critical problems for medical text mining. Medical records which are written by clinicians from different specialties usually contain quite different terminologies and writing styles. The difference of specialties and the cost of human annotation makes it particularly difficult to train a universal medical NER system. In this paper, we propose a labelaware double transfer learning framework (La-DTL) for cross-specialty NER, so that a medical NER system designed for one specialty could be conveniently applied to another one with minimal annotation efforts. The transferability is guaranteed by two components: (i) we propose label-Aware MMD for feature representation transfer, and (ii) we perform parameter transfer with a theoretical upper bound which is also label aware. We conduct extensive experiments on 12 cross-specialty NER tasks. The experimental results demonstrate that La-DTL provides consistent accuracy improvement over strong baselines. Besides, the promising experimental results on non-medical NER scenarios indicate that La-DTL is potential to be seamlessly adapted to a wide range of NER tasks. © 2018 The Association for Computational Linguistics.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

Detecting urban anomalies using multiple spatio-Temporal data sources

引用

Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 2018年第1期2卷 1-18页

作者： Zhang, Huichu Zheng, Yu Yu, Yong Shanghai Jiao Tong University Apex Data and Knowledge Management Lab. 800 Dongchuan Road Shanghai200240 China JD Finance Urban Computing Lab. BDA 18 Kechuang 11 Street 101149 China

Urban anomalies, such as abnormal movements of crowds and accidents, May result in loss of life or property if not handled properly. It would be of great value for governments if anomalies can be automatically alerted in their early stage. However, detecting anomalies in urban area has two main challenges. First, the criteria to determine an anomaly on different occasions (e.g. rainy days vs. sunny days, or holidays vs. workdays) and in different places (e.g. tourist attractions vs. office areas) are distinctly different, as these occasions and places have their own definitions on normal patterns. Second, urban anomalies often exhibit complex forms (e.g. road closure May cause decrease in taxi flow and increase in bike flow). We need an algorithm that not only models the anomaly degree of individual data source but also the combination of changes in multiple data sources. In this paper, we propose a two-step method to tackle those challenges. In the first step, we use a similarity-based algorithm to estimate an anomaly score for each individual data source in each region and time slot based on the values of historically similar regions. Those scores are fed into the second step, where we propose an algorithm based on one-class Support Vector Machine to capture rare patterns occurred in multiple data sources, nearby regions or time slots, and give a final, integrated anomaly score for each region. Evaluations based on both synthetic and real world datasets show the advantages of our method beyond baseline techniques such as distance-based, probability-based methods. 2018 Association for Computing Machinery.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Detection of medical text semantic similarity based on convolutional neural network

引用

BMC MEDICAL INFORMATICS AND DECISION MAKING 2019年第1期19卷 1-11页

作者： Zheng, Tao Gao, Yimei Wang, Fei Fan, Chenhao Fu, Xingzhi Li, Mei Zhang, Ya Zhang, Shaodian Ma, Handong Shanghai Jiao Tong Univ Inst Image Commun & Networking Shanghai Peoples R China Shanghai Jiao Tong Univ Tongren Hosp Shanghai Peoples R China Synyi Res Shanghai Peoples R China Shanghai Jiao Tong Univ APEX Data & Knowledge Management Lab Shanghai Peoples R China Weill Cornell Med Dept Healthcare Policy & Res New York NY USA Shanghai Jiao Tong Univ Shanghai Peoples R China

Background Imaging examinations, such as ultrasonography, magnetic resonance imaging and computed tomography scans, play key roles in healthcare settings. To assess and improve the quality of imaging diagnosis, we need to manually find and compare the pre-existing reports of imaging and pathology examinations which contain overlapping exam body sites from electrical medical records (EMRs). The process of retrieving those reports is time-consuming. In this paper, we propose a convolutional neural network (CNN) based method which can better utilize semantic information contained in report texts to accelerate the retrieving process. Methods We included 16,354 imaging and pathology report-pairs from 1926 patients who admitted to Shanghai Tongren Hospital and had ultrasonic examinations between 1st May 2017 and 31st July 2017. We adapted the CNN model to calculate the similarities among the report-pairs to identify target report-pairs with overlapping body sites, and compared the performance with other six conventional models, including keyword mapping, latent semantic analysis (LSA), latent Dirichlet allocation (LDA), Doc2Vec, Siamese long short term memory (LSTM) and a model based on named entity recognition (NER). We also utilized graph embedding method to enhance the word representation by capturing the semantic relations information from medical ontologies. Additionally, we used LIME algorithm to identify which features (or words) are decisive for the prediction results and improved the model interpretability. Results Experiment results showed that our CNN model gained significant improvement compared to all other conventional models on area under the receiver operating characteristic (AUROC), precision, recall and F1-score in our test dataset. The AUROC of our CNN models gained approximately 3-7% improvement. The AUROC of CNN model with graph-embedding and ontology based medical concept vectors was 0.8% higher than the model with randomly initialized vectors and 1.

关键词： Text similarity Convolutional neural network LIME Natural language processing

来源：评论

学校读者我要写书评

暂无评论

Deep recurrent survival analysis

arXiv

引用

arXiv 2018年

作者： Ren, Kan Qin, Jiarui Zheng, Lei Yang, Zhengyu Zhang, Weinan Qiu, Lin Yu, Yong APEX Data & Knowledge Management Lab Shanghai Jiao Tong University

关键词： Bioinformatics

来源：评论

学校读者我要写书评

暂无评论

label-aware double transfer learning for cross-specialty medical named entity recognition

arXiv

引用

arXiv 2018年

We study the problem of named entity recognition (NER) from electronic medical records, which is one of the most fundamental and critical problems for medical text mining. Medical records which are written by clinicians from different specialties usually contain quite different terminologies and writing styles. The difference of specialties and the cost of human annotation makes it particularly difficult to train a universal medical NER system. In this paper, we propose a label-aware double transfer learning framework (La-DTL) for cross-specialty NER, so that a medical NER system designed for one specialty could be conveniently applied to another one with minimal annotation efforts. The transferability is guaranteed by two components: (i) we propose label-aware MMD for feature representation transfer, and (ii) we perform parameter transfer with a theoretical upper bound which is also label aware. We conduct extensive experiments on 12 cross-specialty NER tasks. The experimental results demonstrate that La-DTL provides consistent accuracy improvement over strong baselines. Besides, the promising experimental results on non-medical NER scenarios indicate that La-DTL is potential to be seamlessly adapted to a wide range of NER tasks. Copyright © 2018, The Authors. All rights reserved.

关键词： Medical computing

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：