Opinions concerning features or aspects of people, entities, products or services are some of the most important textual information. Several methods try to solve the aspect extraction task needed in sentiment analysis by using Deep Learning techniques in specific domains. However, catastrophic forgetting appears when these methods are used to learn aspects of multi-domains. In this paper, we propose a new approach to achieve aspect extraction in multi-domains based on Deep and Lifelong Learning techniques. Our proposal reduces catastrophic forgetting and improves one of the principal state-of-the-art results.
CITATION STYLE
López, D., & Arco, L. (2019). Multi-domain Aspect Extraction Based on Deep and Lifelong Learning. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11896 LNCS, pp. 556–565). Springer. https://doi.org/10.1007/978-3-030-33904-3_52
Mendeley helps you to discover research relevant for your work.