Label Noise Reduction in Entity Typing by Heterogeneous Partial-Label Embedding

By Xiang Ren, Wenqi He, Meng Qu, Clare R. Voss, Heng Ji, Jiawei Han. 2016

149 citations recorded by the source corpus.

Current systems of fine-grained entity typing use distant supervision in conjunction with existing knowledge bases to assign categories (type labels) to entity mentions. However, the type labels so obtained from knowledge bases are often noisy (i.e., incorrect for the entity mention's local context). We define a new task, Label Noise Reduction in Entity Typing (LNR), to be the automatic identification of correct type labels (type-paths) for training examples, given the set of candidate type labels obtained by distant supervision with a given type hierarchy. The unknown type labels for individual entity mentions and the semantic similarity between entity types pose unique challenges for solving the LNR task. We propose a general framework, called PLE, to jointly embed entity mentions, text features and entity types into the same low-dimensional space where, in that space, objects whose types are semantically close have similar representations. Then we estimate the type-path for each training example in a top-down manner using the learned embeddings. We formulate a global objective for learning the embeddings from text corpora and knowledge bases, which adopts a novel margin-based loss that is robust to noisy labels and faithfully models type correlation derived from knowledge bases. Our experiments on three public typing datasets demonstrate the effectiveness and robustness of PLE, with an average of 25% improvement in accuracy compared to next best method.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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 Computer Vision and Pattern Recognition

Citations in the corpus

A Comprehensive Survey of Graph Embedding: Problems, Techniques, and Applications (2017, background): “[84] generate a graph using entity mention, target type and text feature as nodes, and establish three kinds of edges: mention-type, mention-feature and type-type.”
LambdaNet: Probabilistic Type Inference using Graph Neural Networks (2020, methodology): “…might involve looking at object attributes (Farhadi et al., 2017) or label similarity Wang et al. (2018); for natu-ral language, similar techniques are applied to generalize across semantic properties of utterances (Dauphin et al., 2013), entities (Eshel et al., 2017), or labels (Ren et al., 2016).”
Ultra-Fine Entity Typing (2018, methodology): “Overall, as seen in previous work on finegrained NER literature (Gillick et al., 2014; Ren et al., 2016a), finer labels were more challenging to predict than coarse grained labels, and this issue is exacerbated when dealing with ultra-fine types.”
metapath2vec: Scalable Representation Learning for Heterogeneous Networks (2017, methodology): “In addition, several other methods have been proposed for learning representations in networks [4, 5, 11, 20, 23].”
Prompt-Learning for Fine-Grained Entity Typing (2021, methodology): “We follow the version processed by (Ren et al., 2016a), and the data split by (Ren et al., 2016b).”
Label Embedding for Zero-shot Fine-grained Named Entity Typing (2016, methodology): “Current solutions rely on heuristic rules (Gillick et al., 2014) or embedding method (Ren et al., 2016) to remove noises prior to training the multi-label classiﬁer.”
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].”
A Survey on Knowledge Graphs: Representation, Acquisition, and Applications (2020, background): “To reduce label noise, PLE [117] focuses on correct type identiﬁcation and proposes a partial-label embedding model with a heterogeneous graph for the representation of entity mentions, text features, and entity types and their relationships.”