One Vector is Not Enough: Entity-Augmented Distributed Semantics for Discourse Relations

Abstract

Discourse relations bind smaller linguistic units into coherent texts. However, automatically identifying discourse relations is difficult, because it requires understanding the semantics of the linked arguments. A more subtle challenge is that it is not enough to represent the meaning of each argument of a discourse relation, because the relation may depend on links between lower-level components, such as entity mentions. Our solution computes distributional meaning representations by composition up the syntactic parse tree. A key difference from previous work on compositional distributional semantics is that we also compute representations for entity mentions, using a novel downward compositional pass. Discourse relations are predicted from the distributional representations of the arguments, and also of their coreferent entity mentions. The resulting system obtains substantial improvements over the previous state-of-theart in predicting implicit discourse relations in the Penn Discourse Treebank.

Learning to Explicitate Connectives with Seq2Seq Network for Implicit Discourse Relation Classification

“…The first one is denoted as PDTB-Lin (Lin et al, 2009); it uses sections 2-21 for training, 22 as dev and section 23 as test set. The second one is labeled PDTB-Ji (Ji and Eisenstein, 2015), and uses sections 2-20 for training, 0-1 as dev and evaluates on sections 21-22. Our third setting follows the recommendations of , and performs 10-fold cross validation on the whole corpus (sections 0-23).…”

Zero-shot transfer for implicit discourse relation classification

“…The PDTB framework allows annotations to be labelled with more than one label. In such cases we only keep the first label, in line with previous studies (among others Ji and Eisenstein, 2015;Rutherford et al, 2017).…”

“…Recently, neural network approaches have become popular. Ji and Eisenstein (2015) use two RNNs on the syntactic trees of the arguments whereas Zhang et al (2015) use a CNN to perform discourse parsing in a multi-task setting where they consider both explicit and implicit discourse relations. use a simple yet robust feedforward network and achieves the highest performance on the out-of-domain blind test in the CoNLL 2016 shared task .…”

“…Expansion Temporal (Pitler et al, 2009) 21.96 47.13 -16.76 (Zhou et al, 2010) 31.79 47.16 70.11 20.30 (Park and Cardie, 2012) 31.32 49.82 -26.57 (Rutherford and Xue, 2014) 39.70 54.42 70.23 28.69 (Zhang et al, 2015) 33.22 52.04 69.59 30.54 (Ji and Eisenstein, 2015) 35.93 52.78 -27.63 (Lan et al, 2017) 40.73 58.96 72.47 38.50 (Bai and Zhao, 2018) 47.85 54.47 70.60 36.87 (Dai and Huang, 2018) 46 Table 4 use PDTB 3.0 only for training, whereas Table 5 presents the effect of having additional training data from Turkish (a language unrelated to English). Pooling training data from different languages is possible since our model is language-agnostic.…”

Shallow Convolutional Neural Network for Implicit Discourse Relation Recognition

“…Some endto-end deep sentence modeling based approaches have advanced the performance of discourse relation classification. Some researchers (Ji and Eisenstein 2015;Wang and Lan 2016;Qin et al 2017) propose to learn the semantic representation of each argument with neural networks, such as the CNN, the RNN, and the Bi-LSTM for classification.These methods effectively capture inner semantic connections of each argument via distributed representation learning. However, they didn't consider semantic interactions between the two arguments.…”