Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Dec 29;21(Suppl 23):579.
doi: 10.1186/s12859-020-03886-8.

C-Norm: a neural approach to few-shot entity normalization

Arnaud Ferré  1 , Louise Deléger  2 , Robert Bossy  1 , Pierre Zweigenbaum  3 , Claire Nédellec  1
Affiliations

C-Norm: a neural approach to few-shot entity normalization

Arnaud Ferré et al. BMC Bioinformatics. .

Abstract

Background: Entity normalization is an important information extraction task which has gained renewed attention in the last decade, particularly in the biomedical and life science domains. In these domains, and more generally in all specialized domains, this task is still challenging for the latest machine learning-based approaches, which have difficulty handling highly multi-class and few-shot learning problems. To address this issue, we propose C-Norm, a new neural approach which synergistically combines standard and weak supervision, ontological knowledge integration and distributional semantics.

Results: Our approach greatly outperforms all methods evaluated on the Bacteria Biotope datasets of BioNLP Open Shared Tasks 2019, without integrating any manually-designed domain-specific rules.

Conclusions: Our results show that relatively shallow neural network methods can perform well in domains that present highly multi-class and few-shot learning problems.

Keywords: Entity normalization; Few-shot learning; Neural networks; Ontology; Vector space model.

PubMed Disclaimer

Conflict of interest statement

The authors declare that they have no competing interests.

Figures

Fig. 1
Fig. 1
Normalization example for Habitat and Phenotype entities
Fig. 2
Fig. 2
Architecture of the Single Layer Feedforward Neural Network. The input is the matrix of word embeddings of the non-stopword tokens of the mentions
Fig. 3
Fig. 3
Architecture of the shallow CNN. The input is the matrix of word embeddings of the non-stopword tokens of the mentions
Fig. 4
Fig. 4
Architecture of our Sieve method using our SLFNN and shallow CNN methods. The inputs are the matrix of word embeddings of the tokens of the mentions
Fig. 5
Fig. 5
The C-Norm architecture combining the SLFNN and shallow CNN components. The input is the matrix of word embeddings of the tokens of the mentions
Fig. 6
Fig. 6
C-Norm performance scores for Habitats using different decay factors
See this image and copyright information in PMC

References

    1. Faure D, Nédellec C. A corpus-based conceptual clustering method for verb frames and ontology acquisition. In: LREC workshop on adapting lexical and corpus resources to sublanguages and applications. 1998. p. 5–12.
    1. Hwang CH. Incompletely and imprecisely speaking: using dynamic ontologies for representing and retrieving information. KRDB. 1999. p. 13.
    1. Nédellec C, Bossy R, Chaix E, Deleger L. Text-mining and ontologies: new approaches to knowledge discovery of microbial diversity. In: 4th international conference on microbial diversity 2017. Marco Gobetti; 2017.
    1. Bossy R, Chaix E, Deléger L, Ferré A, Ba M, Bessières P, et al. OntoBiotope: une ontologie pour croiser les habitats microbiens avec les analyses de génomes. In: Les journées Bioinformatique de l’INRA. 2016. p. 1.
    1. Ravi S, Larochelle H. Optimization as a model for few-shot learning. In: 8th international conference on learning representations. ICLR 2016, San Juan, Puerto Rico, May 2–4, 2016.

LinkOut - more resources