Filters








4,284 Hits in 4.7 sec

Learning New Facts From Knowledge Bases With Neural Tensor Networks and Semantic Word Vectors [article]

Danqi Chen, Richard Socher, Christopher D. Manning, Andrew Y. Ng
2013 arXiv   pre-print
This model can be improved by initializing entity representations with word vectors learned in an unsupervised fashion from text, and when doing this, existing relations can even be queried for entities  ...  We introduce a neural tensor network (NTN) model which predicts new relationship entries that can be added to the database.  ...  Conclusion We introduced a new model based on Neural Tensor Networks.  ... 
arXiv:1301.3618v2 fatcat:ngou3f56xbddxmaz62sdaltdau

Knowledge Reasoning Based on Neural Tensor Network

Jian-Hui Huang, Jiu-Ming Huang, Ai-Ping Li, Yong-Zhi Tong, L. Long, Y. Li, X. Li, Y. Dai, H. Yang
2017 ITM Web of Conferences  
This paper attempts to explore the model complexity of neural tensor network, a very important method of knowledge reasoning, and the reasoning accuracy.  ...  Knowledge base (KBs) is a very important part of applications such as Q&A system, but the knowledge base is always faced with incompleteness and the lack of inter-entity relationships.  ...  Hence, learning new facts based on the knowledge bases is an essential way to improve them.  ... 
doi:10.1051/itmconf/20171204004 fatcat:jnh7dlb6xfd63jdkuz6lm4euba

Reasoning With Neural Tensor Networks for Knowledge Base Completion

Richard Socher, Danqi Chen, Christopher D. Manning, Andrew Y. Ng
2013 Neural Information Processing Systems  
Lastly, we demonstrate that all models improve when these word vectors are initialized with vectors learned from unsupervised large corpora.  ...  In this paper we introduce an expressive neural tensor network suitable for reasoning over relationships between two entities.  ...  FA8750-13-2-0040, the DARPA Deep Learning program under contract number FA8650-10-C-7020 and NSF IIS-1159679.  ... 
dblp:conf/nips/SocherCMN13 fatcat:nix242ktcvc5zgeje3bya6upsq

Knowledge-Driven Event Embedding for Stock Prediction

Xiao Ding, Yue Zhang, Ting Liu, Junwen Duan
2016 International Conference on Computational Linguistics  
On the other hand, events extracted from raw texts do not contain background knowledge on entities and relations that they are mentioned.  ...  Representing structured events as vectors in continuous space offers a new way for defining dense features for natural language processing (NLP) applications.  ...  2014CB340503, the National Natural Science Foundation of China (NSFC) via Grant 61472107 and 71532004, the Singapore Ministry of Education (MOE) AcRF Tier 2 grant T2MOE201301.  ... 
dblp:conf/coling/DingZLD16 fatcat:cm23wzwyanhvnir3lasyym3ehe

Building Memory with Concept Learning Capabilities from Large-Scale Knowledge Bases

Jiaxin Shi, Jun Zhu
2015 Neural Information Processing Systems  
We present a new perspective on neural knowledge base (KB) embeddings, from which we build a framework that can model symbolic knowledge in the KB together with its learning process.  ...  their embeddings from natural language descriptions, which is very like human's behavior of learning semantic concepts. 1 Concept learning in cognitive science usually refers to the cognitive process  ...  When doing SGD with mini-batch, We back-propagate the error gradients into the neural network, and for CNN, finally into word vectors.  ... 
dblp:conf/nips/ShiZ15 fatcat:f4zozse33ngyddc3wnnote3fsq

Symbolic, Distributed and Distributional Representations for Natural Language Processing in the Era of Deep Learning: a Survey [article]

Lorenzo Ferrone, Fabio Massimo Zanzotto
2019 arXiv   pre-print
A clearer understanding of the strict link between distributed/distributional representations and symbols may certainly lead to radically new deep learning networks.  ...  Recent advances in machine learning (ML) and in natural language processing (NLP) seem to contradict the above intuition: discrete symbols are fading away, erased by vectors or tensors called distributed  ...  Distributed representations are vectors or tensors in metric spaces which underly learning models such as neural networks and also some models based on kernel methods [Zanzotto and Dell'Arciprete 2012a  ... 
arXiv:1702.00764v2 fatcat:xmpga3xn6nhzdos7pcc25z5pde

Logic Tensor Networks: Deep Learning and Logical Reasoning from Data and Knowledge [article]

Luciano Serafini, Artur d'Avila Garcez
2016 arXiv   pre-print
We propose Logic Tensor Networks: a uniform framework for integrating automatic learning and reasoning.  ...  We show how Real Logic can be implemented in deep Tensor Neural Networks with the use of Google's tensorflow primitives.  ...  Approximate satisfiability is defined as a learning task with both knowledge and data being mapped onto real-valued vectors.  ... 
arXiv:1606.04422v2 fatcat:jqttytjqefbdnlktw4pgjkzce4

Compositional Semantics [chapter]

Zhiyuan Liu, Yankai Lin, Maosong Sun
2020 Representation Learning for Natural Language Processing  
After that, we present various typical models for N-ary semantic composition including recurrent neural network, recursive neural network, and convolutional neural network.  ...  Many important applications in NLP fields rely on understanding more complex language units such as phrases, sentences, and documents beyond words.  ...  Based on this philosophy, [15] proposes a recursive neural network to model different levels of semantic units.  ... 
doi:10.1007/978-981-15-5573-2_3 fatcat:uu524rdsxnd7flgrvprsuxaicq

Symbolic, Distributed, and Distributional Representations for Natural Language Processing in the Era of Deep Learning: A Survey

Lorenzo Ferrone, Fabio Massimo Zanzotto
2020 Frontiers in Robotics and AI  
A clearer understanding of the strict link between distributed/distributional representations and symbols may certainly lead to radically new deep learning networks.  ...  Recent advances in machine learning (ML) and in natural language processing (NLP) seem to contradict the above intuition: discrete symbols are fading away, erased by vectors or tensors called distributed  ...  In fact, it is sufficient to generate new Gaussian vectors for new symbols when they appear.  ... 
doi:10.3389/frobt.2019.00153 pmid:33501168 pmcid:PMC7805717 fatcat:353mgx2tr5ftxcx2utx776isou

Building Memory with Concept Learning Capabilities from Large-scale Knowledge Base [article]

Jiaxin Shi, Jun Zhu
2015 arXiv   pre-print
We present a new perspective on neural knowledge base (KB) embeddings, from which we build a framework that can model symbolic knowledge in the KB together with its learning process.  ...  their embeddings from natural language descriptions, which is very like human's behavior of learning semantic concepts.  ...  When doing SGD with mini-batch, We back-propagate the error gradients into the neural network, and for CNN, finally into word vectors.  ... 
arXiv:1512.01173v1 fatcat:xhdhq3fwibd43jgc2qftvlwutu

A Tensor Space Model-Based Deep Neural Network for Text Classification

Han-joon Kim, Pureum Lim
2021 Applied Sciences  
To solve this 'loss of term senses' problem, we develop a concept-driven deep neural network based upon our semantic tensor space model.  ...  This is because a textual document is essentially expressed as a vector (only), albeit with word dimensions, which compromises the inherent semantic information, even if the vector is (appropriately) transformed  ...  Word2Vec, a two-layer neural network with autoencoder characteristics, was trained to learn the relationships among words in a sentence and then found semantic representations [6] .  ... 
doi:10.3390/app11209703 fatcat:s5bp54tdu5djtmuzreehowk5om

A Review of Inference Methods Based on Knowledge Graph [chapter]

Dexiang Zhang, Hairong Wang, Yudan Ding
2020 Frontiers in Artificial Intelligence and Applications  
According to the different methods adopted for each type, each type also includes reasoning based on distributed representation; reasoning based on neural network and mixed reasoning.  ...  Different from traditional knowledge inference methods, knowledge inference methods based on knowledge graphs are also diversified according to their simple, intuitive, flexible and rich knowledge expression  ...  Neural Network Reasoning In single-step reasoning, neural network-based reasoning uses the neural network to directly model the knowledge graph fact tuple, and obtain the vector representation of the fact  ... 
doi:10.3233/faia200727 fatcat:wnjuq5lkbbdh7gzcoukri5nciy

Neural Ranking Models for Document Retrieval [article]

Mohamed Trabelsi, Zhiyu Chen, Brian D. Davison, Jeff Heflin
2021 arXiv   pre-print
A variety of deep learning models have been proposed, and each model presents a set of neural network components to extract features that are used for ranking.  ...  Several approaches to ranking are based on traditional machine learning algorithms using a set of hand-crafted features.  ...  Human activity recognition using recurrent neural networks. In Machine Learning and Knowledge Liu, W., Jia, Y., Sermanet, P., Reed, S.  ... 
arXiv:2102.11903v1 fatcat:zc2otf456rc2hj6b6wpcaaslsa

Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning [article]

Artur d'Avila Garcez, Marco Gori, Luis C. Lamb, Luciano Serafini, Michael Spranger, Son N. Tran
2019 arXiv   pre-print
In spite of the recent impact of AI, several works have identified the need for principled knowledge representation and reasoning mechanisms integrated with deep learning-based systems to provide sound  ...  We illustrate the effectiveness of the approach by outlining the main characteristics of the methodology: principled integration of neural learning with symbolic knowledge representation and reasoning  ...  tensors represent facts (predicates of different arities) from a knowledge base and output tensors represent new facts.  ... 
arXiv:1905.06088v1 fatcat:gm4f3ncukrbevpd7nq5yr75ar4

A Survey on Neural-symbolic Systems [article]

Dongran Yu, Bo Yang, Dayou Liu, Hui Wang
2021 arXiv   pre-print
In this case, an ideal intelligent system--a neural-symbolic system--with high perceptual and cognitive intelligence through powerful learning and reasoning capabilities gains a growing interest in the  ...  Combining the fast computation ability of neural systems and the powerful expression ability of symbolic systems, neural-symbolic systems can perform effective learning and reasoning in multi-domain tasks  ...  The graph is constructed from a knowledge graph and each node is represented by a vector that encodes semantic class information(it is the classes word embedding in this paper).  ... 
arXiv:2111.08164v1 fatcat:bc33afiitnb73bmjtrfbdgkwpy
« Previous Showing results 1 — 15 out of 4,284 results