CIDER: Commonsense Inference for Dialogue Explanation and Reasoning

TL;DR

CIDER introduces a new dataset of dialogue explanations with commonsense knowledge triplets, enabling the development of models that better understand and infer implicit and explicit reasoning in conversations.

Contribution

The paper presents a manually curated dataset of dialogue explanations with annotated commonsense inference triplets, along with three challenging tasks for advancing dialogue understanding.

Findings

Baseline transformer models perform poorly on the tasks.

The dataset covers various types of commonsense knowledge.

Tasks are challenging, indicating room for future research.

Abstract

Commonsense inference to understand and explain human language is a fundamental research problem in natural language processing. Explaining human conversations poses a great challenge as it requires contextual understanding, planning, inference, and several aspects of reasoning including causal, temporal, and commonsense reasoning. In this work, we introduce CIDER -- a manually curated dataset that contains dyadic dialogue explanations in the form of implicit and explicit knowledge triplets inferred using contextual commonsense inference. Extracting such rich explanations from conversations can be conducive to improving several downstream applications. The annotated triplets are categorized by the type of commonsense knowledge present (e.g., causal, conditional, temporal). We set up three different tasks conditioned on the annotated dataset: Dialogue-level Natural Language Inference, Span Extraction, and Multi-choice Span Selection. Baseline results obtained with transformer-based models reveal that the tasks are difficult, paving the way for promising future research. The dataset and the baseline implementations are publicly available at https://cider-task.github.io/cider/.