Fine-Grained Entity Recognition

By Xiao Ling, Daniel S. Weld. AAAI Conference on Artificial Intelligence, 2012

561 citations recorded by the source corpus.

Entity Recognition (ER) is a key component of relation extraction systems and many other natural-language processing applications. Unfortunately, most ER systems are restricted to produce labels from to a small set of entity classes, e.g., person, organization, location or miscellaneous. In order to intelligently understand text and extract a wide range of information, it is useful to more precisely determine the semantic classes of entities mentioned in unstructured text. This paper defines a fine-grained set of 112 tags, formulates the tagging problem as multi-class, multi-label classification, describes an unsupervised method for collecting training data, and presents the FIGER implementation. Experiments show that the system accurately predicts the tags for entities. Moreover, it provides useful information for a relation extraction system, increasing the F1 score by 93%. We make FIGER and its data available as a resource for future work.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/777ef98b19cac4dd83bd2318f6d29a38ea6c8294

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

Citations in the corpus

CoType: Joint Extraction of Typed Entities and Relations with Knowledge Bases (2016, methodology): “To test the performance on entity recognition and typing, we also compare with several entity recognition systems, including a supervised method HYENA [55], distant supervision methods (FIGER [26], Google [15], WSABIE [54]), and a noise-robust approach PLE [42].”
Entity Linking via Joint Encoding of Types, Descriptions, and Context (2017, methodology): “Our performance is competitive with FIGER (Ling and Weld, 2012) and neural-LSTM model of Shimaoka et al.”
On the Importance of Delexicalization for Fact Verification (2019, background): “More importantly, this also suggests that while we have demonstrated that semantic masking is clearly ben-eﬁcial, it is unclear exactly what level of masking granularity should be employed for a given domain – from coarse-grained NER tags, to the more granular super sense tags, perhaps even to very ”
Learning Cross-Context Entity Representations from Text (2020, background): “Mention-level entity typing Another well studied task is mention-level entity typing (e.g. Ling & Weld, 2012; Choi et al., 2018).”
Few-NERD: A Few-shot Named Entity Recognition Dataset (2021, methodology): “3200 entity annotation, which is inspired and modified from previous fine-grained entity recognition studies (Ling and Weld, 2012; Gillick et al., 2014; Choi et al., 2018; Ringland et al., 2019).”
Cross-lingual Name Tagging and Linking for 282 Languages (2017, background): “Many fine-grained entity typing approaches (Fleischman and Hovy, 2002; Giuliano, 2009; Ekbal et al., 2010; Ling and Weld, 2012; Yosef et al., 2012; Nakashole et al., 2013; Gillick et al., 2014; Yogatama et al., 2015; Del Corro et al., 2015) also created annotations based on Wikipedia anchor links.”
Pretrained Encyclopedia: Weakly Supervised Knowledge-Pretrained Language Model (2019, methodology): “To compare with an existing study (Zhang et al., 2019) that also attempts to incorporate entity knowledge into language models, we consider an additional entity typing task using the large FIGER dataset (Ling & Weld, 2012).”
Improving Entity Linking by Modeling Latent Entity Type Information (2020, methodology): “(2018) consisting of more than 10,000 ultra-fine grained types; the other one is the FIGER type sets (Ling and Weld 2012) consisting of 112 fine grained types.”
Generalizing to Unseen Entities and Entity Pairs with Row-less Universal Schema (2016, unknown): “We simply sample uniformly from the entire set of row entries Entity type prediction at the individual sentence level has been studied extensively (Pantel et al., 2012; Ling and Weld, 2012; Shimaoka et al., 2016).”
K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters (2020, methodology): “We adopt two metrics including ExactMatch (EM) and loose F1 (Ling and Weld, 2012) scores to evaluate our models.”
Prompt-Learning for Fine-Grained Entity Typing (2021, methodology): “In terms of evaluation metrics, we follow the widely used setting of Ling and Weld (2012), which includes strict accuracy (Acc), loose macro F1-score (MaF) and loose micro F1-score (MiF) to evaluate the performances of models.”
Label Noise Reduction in Entity Typing by Heterogeneous Partial-Label Embedding (2016, methodology): “Many previous studies ignore the label noise in automatically labeled training corpora— all candidate types obtained by distant supervision are treated as “true” types in training multi-label (hierarchical) classiﬁers [34, 14, 35].”
Mention and Entity Description Co-Attention for Entity Disambiguation (2018, methodology): “Following (Ling and Weld 2012), we filter out irrelevant types to reduce the data noise, and only keep the well-maintained types (e.”
Entity Linking with a Knowledge Base: Issues, Techniques, and Solutions (2015, background): “For unlinkable mentions, there are some studies that identify their fine-grained types from the knowledge base [12], [13], [14], [15], which is out of scope for entity linking systems.”
Knowledge Graphs (2020, unknown): “Distant supervision [319, 436, 572] uses known entities in a knowledge graph as seed examples through which similar entities can be detected.”
Redcoat: A Collaborative Annotation Tool for Hierarchical Entity Typing (2019, background): “Redcoat also features three category hierarchy presets: NER , the standard Named Entity Recognition classes ( PER , LOC , ORG , MISC ), FIGER (Ling and Weld, 2012) (ﬁne-grained entity recognition), and Mining , containing categories speciﬁc to workplace accident data in the mining industry.”
Embedding Methods for Fine Grained Entity Type Classification (2015, methodology): “F IGER evaluation Our second evaluation dataset is F IGER from Ling and Weld (2012).”