2016 ComposingDistributedRepresentat
- (Takase et al., 2016) ⇒ Sho Takase, Naoaki Okazaki, and Kentaro Inui. (2016). “Composing Distributed Representations of Relational Patterns.” In: Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016). doi:DOI:10.18653/v1/p16-1215 arXiv:1707.07265
Subject Headings: Relational Pattern; Relational Pattern Learning System.
Notes
Cited By
- Google Scholar: ~ 10 Citations
- Semantic Scholar: ~ 8 Citations
Quotes
Abstract
Learning distributed representations for relation instances is a central technique in downstream NLP applications. In order to address semantic modeling of relational patterns, this paper constructs a new dataset that provides multiple similarity ratings for every pair of relational patterns on the existing dataset. In addition, we conduct a comparative study of different encoders including additive composition, RNN, LSTM, and GRU for composing distributed representations of relational patterns. We also present Gated Additive Composition, which is an enhancement of additive composition with the gating mechanism. Experiments show that the new dataset does not only enable detailed analyses of the different encoders, but also provides a gauge to predict successes of distributed representations of relational patterns in the relation classification task.
1 Introduction
In particular, semantic modeling of relations and their textual realizations (relational patterns hereafter) is extremely important because a relation (e.g., causality) can be mentioned by various expressions (e.g., “X cause Y”, “X lead to Y”, “Y is associated with X”). To make matters worse, relational patterns are highly productive: we can produce a emphasized causality pattern “X increase the severe risk of Y” from “X increase the risk of Y” by inserting severe to the pattern. To model the meanings of relational patterns, the previous studies built a co-occurrence matrix between relational patterns (e.g., “X increase the risk of Y”) and entity pairs (e.g., “X: smoking, Y: cancer”) (Lin and Pantel, 2001; Nakashole et al., 2012). Based on the distributional hypothesis (Harris, 1954), we can compute a semantic vector of a relational pattern from the co-occurrence matrix, and measure the similarity of two relational patterns as the cosine similarity of the vectors.
References
;
Author | volume | Date Value | title | type | journal | titleUrl | doi | note | year | |
---|---|---|---|---|---|---|---|---|---|---|
2016 ComposingDistributedRepresentat | Kentaro Inui Naoaki Okazaki Sho Takase | Composing Distributed Representations of Relational Patterns | DOI:10.18653/v1/p16-1215 | 2016 |