Commonsense-Focused Dialogues for Response Generation: An Empirical Study

By Pei Zhou, Karthik Gopalakrishnan, Behnam Hedayatnia, Seokhwan Kim, Jay Pujara, Xiang Ren, Yang Liu, Dilek Z. Hakkani-Tür. 2021

50 citations recorded by the source corpus.

Smooth and effective communication requires the ability to perform latent or explicit commonsense inference. Prior commonsense reasoning benchmarks (such as SocialIQA and CommonsenseQA) mainly focus on the discriminative task of choosing the right answer from a set of candidates, and do not involve interactive language generation as in dialogue. Moreover, existing dialogue datasets do not explicitly focus on exhibiting commonsense as a facet. In this paper, we present an empirical study of commonsense in dialogue response generation. We first auto-extract commonsensical dialogues from existing dialogue datasets by leveraging ConceptNet, a commonsense knowledge graph. Furthermore, building on social contexts/situations in SocialIQA, we collect a new dialogue dataset with 25K dialogues aimed at exhibiting social commonsense in an interactive setting. We evaluate response generation models trained using these datasets and find that models trained on both extracted and our collected data produce responses that consistently exhibit more commonsense than baselines. Finally we propose an approach for automatic evaluation of commonsense that relies on features derived from ConceptNet and pre-trained language and dialog models, and show reasonable correlation with human evaluation of responses’ commonsense quality.

Subfields: Artificial Intelligence, General Health Professions, Natural Language Processing, Dialogue & Conversational AI

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In General Health Professions

Dilek Z. Hakkani-Tür

In Natural Language Processing

In Dialogue & Conversational AI

Citations in the corpus

Lawyers are Dishonest? Quantifying Representational Harms in Commonsense Knowledge Resources (2021, background): “Numerous models have included knowledge resources such as ConceptNet (Speer et al., 2017) for question answering (Lin et al., 2019), sarcasm generation (Chakrabarty et al., 2020), and dialogue response generation (Zhou et al., 2018, 2021), among others.”
Think Before You Speak: Explicitly Generating Implicit Commonsense Knowledge for Response Generation (2021, methodology): “For training, we use the filtered version of the four datasets from Zhou et al. (2021a), which ensures each dialogue contains at least one commonsense knowledge triple from ConceptNet.”
Think Before You Speak: Using Self-talk to Generate Implicit Commonsense Knowledge for Response Generation (2021, methodology): “We use dialogue datasets from Zhou et al. (2021a) since they proposed “ commonsense-focused dialogues ” by ﬁltering three existing dialogue datasets DailyDialog (Li et al., 2017), EmpatheticDia-logues (Rashkin et al., 2019), MuTual (Cui et al., 2020) using ConceptNet triples, and also crowd-sourced…”