Context-Aware Representations for Knowledge Base Relation Extraction

By Daniil Sorokin, Iryna Gurevych. 2017

167 citations recorded by the source corpus.

We demonstrate that for sentence-level relation extraction it is beneficial to consider other relations in the sentential context while predicting the target relation. Our architecture uses an LSTM-based encoder to jointly learn representations for all relations in a single sentence. We combine the context representations with an attention mechanism to make the final prediction. We use the Wikidata knowledge base to construct a dataset of multiple relations per sentence and to evaluate our approach. Compared to a baseline system, our method results in an average error reduction of 24 on a held-out set of relations. The code and the dataset to replicate the experiments are made available at https://github.com/ukplab/.

Subfields: Artificial Intelligence, Information Extraction, Knowledge Graphs & Representation

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Extraction

In Knowledge Graphs & Representation

Citations in the corpus

Learning How to Ask: Querying LMs with Mixtures of Soft Prompts (2021, methodology): “Factual knowledge is traditionally extracted from large corpora using a pipeline of NLP tools (Surdeanu and Ji, 2014), including entity extraction (Lample et al., 2016), entity linking (Rao et al., 2013) and relation extraction (Sorokin and Gurevych, 2017).”
Eliciting Knowledge from Language Models Using Automatically Generated Prompts (2020, methodology): “We compare A UTO P ROMPT to LAMA and LPAQA (their prompts are still useful here), as well as a recent supervised relation extraction model (Sorokin and Gurevych, 2017) that was also used by Petroni et al. (2019).”
Assessing The Factual Accuracy of Generated Text (2019, methodology): “multiple entities, we also add a prefixSUBJfor ei and OBJto ej as a hint. For example, X = Person1 was born in City1 becomes Sl = SUBJ{ Person1 } was born in OBJ{ City1 }. Unlike Sorokin and Gurevych [32], our classifier does not require additional context. Let si be a token in the input sentence Sl”
Reasoning with Latent Structure Refinement for Document-Level Relation Extraction (2020, background): “…neural networks (CNN) (Zeng et al., 2014), LSTM, bidirectional LSTM (BiLSTM) (Cai et al., 2016) and attention-based LSTM (Con-textAware) (Sorokin and Gurevych, 2017 As shown in Table 2, LSR with GloVe achieves 54.18 F 1 on the test set, which is the new state-of-the-art result for…”
Coreferential Reasoning Learning for Language Representation (2020, background): “(2) ContextAware (Sorokin and Gurevych, 2017) takes relations’ interaction into account, which demonstrates that other relations in the context are beneficial for target relation prediction.”
Extracting Multiple-Relations in One-Pass with Pre-Trained Transformers (2019, methodology): “Existing MRE approaches are based on either feature and model architecture selection techniques (Xu et al., 2015; Gormley et al., 2015; Nguyen and Grishman, 2015; F. Petroni and Gemulla, 2015; Sorokin and Gurevych, 2017; Song et al., 2018b), or domain adaptations approaches (Fu et al.”
SemEval-2020 Task 6: Definition Extraction from Free Text with the DEFT Corpus (2020, background): “er of additional possible tags in the DEFT corpus. Given the precedence of previous performance on other more constrained relation extraction tasks (see also Lin et al. (2016), Shen and Huang (2016), Sorokin and Gurevych (2017), etc.), we predicted this task would see mostly low performing scores. H”
DocRED: A Large-Scale Document-Level Relation Extraction Dataset (2019, methodology): “As a result, RE systems that model context information with algorithms of quadratic or even higher computational complexity such as (Sorokin and Gurevych, 2017; Christopoulou et al., 2018) are not efﬁcient enough for document-level RE.”