Social Commonsense Reasoning with Multi-Head Knowledge Attention
This work addresses the problem of enhancing AI's ability to reason about social contexts for applications in natural language understanding, though it appears incremental as it builds on existing transformer-based methods.
The paper tackled social commonsense reasoning by proposing a multi-head knowledge attention model that encodes commonsense inference rules, improving performance over state-of-the-art models like RoBERTa on tasks such as Abductive Natural Language Inference and Counterfactual Invariance Prediction, with the model being the first to show that counterfactual reasoning aids in abductive reasoning.
Social Commonsense Reasoning requires understanding of text, knowledge about social events and their pragmatic implications, as well as commonsense reasoning skills. In this work we propose a novel multi-head knowledge attention model that encodes semi-structured commonsense inference rules and learns to incorporate them in a transformer-based reasoning cell. We assess the model's performance on two tasks that require different reasoning skills: Abductive Natural Language Inference and Counterfactual Invariance Prediction as a new task. We show that our proposed model improves performance over strong state-of-the-art models (i.e., RoBERTa) across both reasoning tasks. Notably we are, to the best of our knowledge, the first to demonstrate that a model that learns to perform counterfactual reasoning helps predicting the best explanation in an abductive reasoning task. We validate the robustness of the model's reasoning capabilities by perturbing the knowledge and provide qualitative analysis on the model's knowledge incorporation capabilities.