我们的目标是将符号逻辑推理丰富的多步推理与神经网络的泛化能力结合起来。我们对文本和大规模知识库(KBs)中的实体和关系的复杂推理特别感兴趣。Neelakantan et al.(2015)使用RNN来构建KB中多跳路径的分布式语义;然而,由于多种原因,该方法缺乏准确性和实用性。本文提出了三个重要的建模进展:(1)我们学会了联合推理关系,实体和实体类型;(2)我们使用神经注意力建模来整合多个路径;(3)我们学会了在一个单一的RNN中共享力量,该RNN代表所有关系的逻辑组成。在一个大规模的Freebase+ Cocktail Web预测任务,我们实现了25%的错误减少,和53%的错误减少稀疏的关系,由于共享的力量。在WordNet中的推理链上,我们将平均分位数的错误减少了84%,与以前的最先进技术相比。代码和数据可在https://rajarshd.github.io/ChainsofReasoning上获得
英文摘要:
Our goal is to combine the rich multistep inference of symbolic logical reasoning with the generalization capabilities of neural networks. We are particularly interested in complex reasoning about entities and relations in text and large-scale knowledge bases (KBs). Neelakantan et al. (2015) use RNNs to compose the distributed semantics of multi-hop paths in KBs; however for multiple reasons, the approach lacks accuracy and practicality. This paper proposes three significant modeling advances: (1) we learn to jointly reason about relations, entities, and entity-types; (2) we use neural attention modeling to incorporate multiple paths; (3) we learn to share strength in a single RNN that represents logical composition across all relations. On a largescale Freebase+ClueWeb prediction task, we achieve 25% error reduction, and a 53% error reduction on sparse relations due to shared strength. On chains of reasoning in WordNet we reduce error in mean quantile by 84% versus previous state-of-the-art. The code and data are available at https://rajarshd.github.io/ChainsofReasoning
Learning the grammar of drug prescription: recurrent neural network grammars for medication information extraction in clinical texts
CHU Necker, AP-HP(巴黎公共医疗救助机构内克尔大学医院);INSERM UMR-S 872(法国国家健康与医学研究院 UMR-S 872);Sorbonne Paris Cité, Faculté de Médecine(索邦巴黎西岱大学医学院);Université de Paris(巴黎大学)
Terminologies augmented recurrent neural network model for clinical named entity recognition
Paris University(巴黎大学);AP-HP(巴黎公共援助医院);LIMSI(法国计算机科学与自动化研究所);CNRS(法国国家科学研究中心);Univ. Paris-Sud(巴黎南大学);Université Paris-Saclay(巴黎萨克雷大学);Sorbonne Université(索邦大学);Inserm(法国国家健康与医学研究院);Univ Paris 13(巴黎第十三大学);LIMICS(医学信息学与电子健康知识工程实验室)
我们的目的是提高性能的监督模型的临床命名实体识别(NER)使用医学术语。为了评估我们的系统在法国,我们建立了一个语料库的5种类型的临床实体。我们使用了一个基于术语的系统作为基线,建立在UMLS和SNOMED。然后,我们评估了一个biGRU-CRF,和一个混合系统使用的预测的术语为基础的系统的biGRU-CRF的功能。在英语中,我们在i2 b2 -2009药物名称识别的药物挑战赛上评估了NER系统,其中包含268个文档的8,573个实体。在法语中,我们建立了APcNER,这是一个包含147个文档的语料库,注释了5个实体(药物名称,体征或症状,疾病或病症,诊断程序或实验室测试和治疗程序)。我们评估每个NER系统使用精确和部分匹配定义的F-测度的NER。APcNER包含4,837个实体,需要28小时进行注释,注释者间一致性对于精确匹配的药物名称是可接受的(85%),对于非精确匹配的其他实体类型是可接受的(>70%)。对于i2 b2 -2009和APcNER上的药物名称识别,biGRU-CRF比基于术语的系统表现更好,精确匹配F测量值分别为91.1%和73%,81.9%和75%。此外,混合系统优于biGRU-CRF,精确匹配F测量值为92.2% vs 91.1%(i2 b2 -2009)和88.4% vs 81.9%(APcNER)。在APcNER语料库上,混合系统对5个实体的微平均F-测…
DARER: Dual-task Temporal Relational Recurrent Reasoning Network for Joint Dialog Sentiment Classification and Act Recognition
University of Technology Sydney(悉尼科技大学);Australian Artificial Intelligence Institute(澳大利亚人工智能研究所);Centre for Frontier Artificial Intelligence Research(前沿人工智能研究中心)
One Subgraph for All: Efficient Reasoning on Opening Subgraphs for Inductive Knowledge Graph Completion
School of Computer Science, Central China Normal University(华中师范大学计算机学院);Faculty of Artificial Intelligence in Education, Central China Normal University(华中师范大学人工智能教育学部);School of Computer Science, Wuhan University(武汉大学计算机学院);York University(约克大学)
KnowRA: Knowledge Retrieval Augmented Method for Document-level Relation Extraction with Comprehensive Reasoning Abilities
Nanjing Normal University, School of Computer Science and Electronic Informatics(南京师范大学计算机科学与电子信息学院);State Key Laboratory for Novel Software Technology at Nanjing University(南京大学新型软件技术国家重点实验室);Nanjing University, School of Computer Science(南京大学计算机科学学院)
UniKGQA: Unified Retrieval and Reasoning for Solving Multi-hop Question Answering Over Knowledge Graph
Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学高瓴人工智能学院);School of Information, Renmin University of China(中国人民大学信息学院);Beijing Key Laboratory of Big Data Management and Analysis Methods(大数据管理与分析方法北京市重点实验室)
Expanding the Scope: Inductive Knowledge Graph Reasoning with Multi-Starting Progressive Propagation
State Key Laboratory for Novel Software Technology, Nanjing University(南京大学计算机软件新技术国家重点实验室);National Institute of Healthcare Data Science, Nanjing University(南京大学健康医疗大数据国家研究院)
ClueReader: Heterogeneous Graph Attention Network for Multi-hop Machine Reading Comprehension
Qufu Normal University(曲阜师范大学);East China Normal University(华东师范大学);Shanghai Institute of AI for Education(上海人工智能教育研究院);Harbin Institute of Technology(哈尔滨工业大学);HUTECH University(胡志明市工业大学);Andalusian Research Institute in Data Science and Computational Intelligence (DaSCI)(安达卢西亚数据科学与计算智能研究院(DaSCI));University of Granada(格拉纳达大学);Iwate Prefectural University(岩手县立大学)
Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach
ELLIS Institute Tübingen(埃利斯研究所蒂宾根分所);Max-Planck Institute for Intelligent Systems(马克斯·普朗克智能系统研究所);Tübingen AI Center(蒂宾根人工智能中心);University of Maryland, College Park(马里兰大学帕克分校);Lawrence Livermore National Laboratory(劳伦斯利弗莫尔国家实验室)
Depth-Recurrent Attention Mixture (Dreamer): Giving Latent Reasoning the Attention it Deserves
Aleph Alpha Research, Heidelberg, Germany(Aleph Alpha研究机构);Technical University of Munich, Germany(慕尼黑技术大学);Research Group Social Computing, Technical University of Munich, Germany(社会计算研究组,慕尼黑技术大学);Lab1141, Germany(Lab1141研究机构);Computer Science Department, TU Darmstadt, Germany(图林根达姆施塔特技术大学计算机科学系)
Dynamic Anticipation and Completion for Multi-Hop Reasoning over Sparse Knowledge Graph
Department of Computer Science and Technology, BNRist(计算机科学与技术系,BNRist);KIRC, Institute for Artificial Intelligence, Tsinghua University(清华大学人工智能研究院,知识图谱与智能推理中心(KIRC));Alibaba Group(阿里巴巴集团)