Alert button

Implicit Premise Generation with Discourse-aware Commonsense Knowledge Models

Sep 11, 2021
Tuhin Chakrabarty, Aadit Trivedi, Smaranda Muresan

Figure 1 for Implicit Premise Generation with Discourse-aware Commonsense Knowledge Models
Figure 2 for Implicit Premise Generation with Discourse-aware Commonsense Knowledge Models
Figure 3 for Implicit Premise Generation with Discourse-aware Commonsense Knowledge Models
Figure 4 for Implicit Premise Generation with Discourse-aware Commonsense Knowledge Models

Share this with someone who'll enjoy it:

Enthymemes are defined as arguments where a premise or conclusion is left implicit. We tackle the task of generating the implicit premise in an enthymeme, which requires not only an understanding of the stated conclusion and premise but also additional inferences that could depend on commonsense knowledge. The largest available dataset for enthymemes (Habernal et al., 2018) consists of 1.7k samples, which is not large enough to train a neural text generation model. To address this issue, we take advantage of a similar task and dataset: Abductive reasoning in narrative text (Bhagavatula et al., 2020). However, we show that simply using a state-of-the-art seq2seq model fine-tuned on this data might not generate meaningful implicit premises associated with the given enthymemes. We demonstrate that encoding discourse-aware commonsense during fine-tuning improves the quality of the generated implicit premises and outperforms all other baselines both in automatic and human evaluations on three different datasets.

* EMNLP 2021 Camera ready  
View paper onarxiv icon

Share this with someone who'll enjoy it: