• Corpus ID: 14344787

Extracting Multilingual Relations under Limited Resources: TAC 2016 Cold-Start KB construction and Slot-Filling using Compositional Universal Schema

  title={Extracting Multilingual Relations under Limited Resources: TAC 2016 Cold-Start KB construction and Slot-Filling using Compositional Universal Schema},
  author={Haw-Shiuan Chang and Abdurrahman Munir and Ao Liu and Johnny Tian-Zheng Wei and Aaron Traylor and Ajay Nagesh and Nicholas Monath and Pat Verga and Emma Strubell and Andrew McCallum},
  journal={Theory and Applications of Categories},
We describe the UMass IESL relation extraction system for TAC KBP 2016. One of the main challenges in TAC 2016 is to extract relations from multiple languages, including those with relatively low resources like Spanish. To mitigate the problem, we integrate multilingual and compositional universal schema from Verga et al. (2016) into our slot filling and knowledge base construction pipelines. The flexibility of our universal schema framework allows us to extract high quality Spanish relations… 

Figures and Tables from this paper

Inducing Implicit Relations from Text Using Distantly Supervised Deep Nets

An extensive evaluation of the system across three different benchmarks with different characteristics is provided, showing that the unified framework can consistently outperform state of the art solutions.

Deep learning methods for knowledge base population

This thesis explores deep learning methods for automatically populating knowledge bases from text, addressing the following tasks: slot filling, uncertainty detection and type-aware relation extraction, and is the first to integrate an uncertainty detection component into a slot filling pipeline.

Multi-facet Universal Schema

This work proposes multi-facet universal schema that uses a neural model to represent each sentence pattern as multiple facet embeddings and encourages one of these facet embedDings to be close to that of another sentence pattern if they co-occur with the same entity pair.

Type-aware Convolutional Neural Networks for Slot Filling

This paper proposes type-aware convolutional neural networks to benefit from the mutual dependencies between entity and relation classification, and explores different ways of integrating the named entity types of the relation arguments into a neural network for relation classification.

MIDAS: Finding the Right Web Sources to Fill Knowledge Gaps

MIDAS is presented, a system that harnesses the results of automated knowledge extraction pipelines to repair the bottleneck in industrial knowledge creation and augmentation processes and introduces a novel concept, web source slices, to describe the contents of a web source.

Knowledge Base Population based on Entity Graph Analysis. (Peuplement d'une base de connaissance fondé sur l'exploitation d'un graphe d'entités)

This work proposes to extract a graph of entities from the overall corpus and to compute features on this graph that are able to capture some evidence of holding relationships between a pair of entities and tests these features on a task of relation validation.

Stacking With Auxiliary Features

This paper proposes stacking with auxiliary features that learns to fuse relevant information from multiple systems to improve performance and demonstrates the power and generality of this approach on three very different and difficult problems.

Improving and understanding data quality in large-scale data systems

This research presents a meta-modelling framework that automates the very labor-intensive and therefore time-heavy and expensive process of manually cataloging and evaluating data quality in a large-scale data system.

Discovering Implicit Knowledge with Unary Relations

A new methodology to identify relations between two entities, consisting of detecting a very large number of unary relations, and using them to infer missing entities, is proposed by describing a deep learning architecture able to learn thousands of such relations very efficiently by using a common deep learning based representation.



Multilingual Relation Extraction using Compositional Universal Schema

This paper's model outperforms the top system from TAC 2013 slot-filling using no handwritten patterns or additional annotation, and is suited to broad-coverage automated knowledge base construction in a variety of languages and domains.

Relation Extraction with Matrix Factorization and Universal Schemas

This work presents matrix factorization models that learn latent feature vectors for entity tuples and relations that achieve substantially higher accuracy than a traditional classification approach and is able to reason about unstructured and structured data in mutually-supporting ways.

Lexicon Infused Phrase Embeddings for Named Entity Resolution

A new form of learning word embeddings that can leverage information from relevant lexicons to improve the representations, and the first system to use neural word embedDings to achieve state-of-the-art results on named-entity recognition in both CoNLL and Ontonotes NER are presented.

Joint Learning of the Embedding of Words and Entities for Named Entity Disambiguation

A novel embedding method specifically designed for NED that jointly maps words and entities into the same continuous vector space and extends the skip-gram model by using two models.

Capturing Semantic Similarity for Entity Linking with Convolutional Neural Networks

This work presents a model that uses convolutional neural networks to capture semantic correspondence between a mention's context and a proposed target entity, outperforming the prior systems of Durrett and Klein (2014) and Nguyen et al. (2014).

Leveraging Deep Neural Networks and Knowledge Graphs for Entity Disambiguation

A novel deep semantic relatedness model (DSRM) based on deep neural networks (DNN) and semantic knowledge graphs (KGs) to measure entity semantic relatednesses for topical coherence modeling is presented.

RelationFactory: A Fast, Modular and Effective System for Knowledge Base Population

RelationFactory emphasizes modularity, is easily configurable and uses a transparent pipelined approach and a recall error analysis component categorizes and illustrates cases in which the system missed a correct answer.

Comparing Convolutional Neural Networks to Traditional Models for Slot Filling

A convolutional neural network is proposed which splits the input sentence into three parts according to the relation arguments and is compared to state-ofthe-art and traditional approaches of relation classification.

Modeling Mention, Context and Entity with Neural Networks for Entity Disambiguation

A new neural network approach is presented that takes consideration of the semantic representations of mention, context and entity, encodes them in continuous vector space and effectively leverages them for entity disambiguation.

Joint Learning of Local and Global Features for Entity Linking via Neural Networks

This work introduces a novel framework based on convolutional neural networks and recurrent neural networks to simultaneously model the local and global features for entity linking and examines the entity linking systems on the domain adaptation setting that further demonstrates the cross-domain robustness of the proposed model.