RECORD
Neural Metric Learning for Fast End-to-End Relation Extraction
Computation and Language · Submitted 17 May 2019
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim We validate our model on the ADE and CoNLL04 datasets for end-to-end RE and demonstrate $\approx 1\%$ gain (in F-score) over prior best results with training and testing times that are seven to ten times faster --- the latter highly advantageous for time-sensitive end user applications.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Relation extraction (RE) is an indispensable information extraction task in several disciplines. RE models typically assume that named entity recognition (NER) is already performed in a previous step by another independent model. Several recent efforts, under the theme of end-to-end RE, seek to exploit inter-task correlations by modeling both NER and RE tasks jointly. Earlier work in this area commonly reduces the task to a table-filling problem wherein an additional expensive decoding step involving beam search is applied to obtain globally consistent cell labels. In efforts that do not employ table-filling, global optimization in the form of CRFs with Viterbi decoding for the NER component is still necessary for competitive performance. We introduce a novel neural architecture utilizing the table structure, based on repeated applications of 2D convolutions for pooling local dependency and metric-based features, that improves on the state-of-the-art without the need for global optimization. We validate our model on the ADE and CoNLL04 datasets for end-to-end RE and demonstrate $\approx 1\%$ gain (in F-score) over prior best results with training and testing times that are seven to ten times faster --- the latter highly advantageous for time-sensitive end user applications.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Relation extraction (RE) is an indispensable information extraction task in several disciplines. | background 1.00 | 0.04 | 0.02 | asserted 1.00 |
| s1 | RE models typically assume that named entity recognition (NER) is already performed in a previous step by another independent model. | background 0.94 | 0.04 | 0.02 | asserted 1.00 |
| s2 | Several recent efforts, under the theme of end-to-end RE, seek to exploit inter-task correlations by modeling both NER and RE tasks jointly. | background 1.00 | 0.05 | 0.03 | asserted 1.00 |
| s3 | Earlier work in this area commonly reduces the task to a table-filling problem wherein an additional expensive decoding step involving beam search is applied to obtain globally consistent cell labels. | background 0.96 | 0.04 | 0.03 | asserted 1.00 |
| s4 | In efforts that do not employ table-filling, global optimization in the form of CRFs with Viterbi decoding for the NER component is still necessary for competitive performance. | background 0.93 | 0.05 | 0.03 | asserted 0.99 |
| s5 | We introduce a novel neural architecture utilizing the table structure, based on repeated applications of 2D convolutions for pooling local dependency and metric-based features, that improves on the state-of-the-art without the need for global optimization. | contribution 0.66 | 0.94 | 0.62 | asserted 1.00 |
| s6 | We validate our model on the ADE and CoNLL04 datasets for end-to-end RE and demonstrate $\approx 1\%$ gain (in F-score) over prior best results with training and testing times that are seven to ten times faster --- the latter highly advantageous for time-sensitive end user applications. | result 1.00 | 0.97 | 0.70 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v4
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Relation extraction (RE) is an indispensable information extraction task in several disciplines. RE models typically assume that named entity recognition (NER) is already performed in a previous step by another independent model. Several recent efforts, under the theme of end-to-end RE, seek to exploit inter-task correlations by modeling both NER and RE tasks jointly. Earlier work in this area commonly reduces the task to a table-filling problem wherein an additional expensive decoding step involving beam search is applied to obtain globally consistent cell labels. In efforts that do not employ table-filling, global optimization in the form of CRFs with Viterbi decoding for the NER component is still necessary for competitive performance. We introduce a novel neural architecture utilizing the table structure, based on repeated applications of 2D convolutions for pooling local dependency and metric-based features, that improves on the state-of-the-art without the need for global optimization. We validate our model on the ADE and CoNLL04 datasets for end-to-end RE and demonstrate $\approx 1\%$ gain (in F-score) over prior best results with training and testing times that are seven to ten times faster --- the latter highly advantageous for time-sensitive end user applications.
Human review
How the literature uses it
7 arXiv papers cite this in 8 sentences
Papers citing it · 2 sentences without a confident label
Compares with it 2 papers
-
Moreover, we also compare to the baselines that achieve competitive performance, including Multi-head [ref] , Multi-head + AT [ref] , Rel-Metric [ref] , SpERT [ref] .
Global Pointer: Novel Efficient Span-based Approach for Named Entity Recognition · Experimental Setup · 1.00
-
To further highlight the validity of our model in the general field, we use a variety of complex neural arrest models as the baseline (multi-head+AT [ref] , multi-head [ref] , Relation-Metric with AT [ref] , Relation-Metric with AT [ref] , Spert [ref] ).
AutoIE: An Automated Framework for Information Extraction from Scientific Literature · Baseline and Parameter Set · 0.81
Cites it as background 3 papers
-
(2) A joint modeling approach that simultaneously extracts both entities and relations [ref] .
End-to-End $n$-ary Relation Extraction for Combination Drug Therapies · Related Work · 0.97
-
Specifically, some of them cast NER and RE as a joint table filling problem [ref] .
Towards Effective Multi-Task Interaction for Entity-Relation Extraction: A Unified Framework with Selection Recurrent Network · Introduction · 0.81
-
Early RE research applied approximate frequency counting and dimension reduction to speed up similarity computation in unsupervised RE [ref] , and adopted neural metric learning methods to accelerate RE [ref] .
FastRE: Towards Fast Relation Extraction with Convolutional Encoder and Improved Cascade Binary Tagging Framework · Introduction · 0.92
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.