Enhancing Natural Language Representation with Large-Scale Out-of-Domain Commonsense
This work addresses the challenge of effectively using commonsense knowledge in natural language processing tasks, offering a solution that avoids catastrophic forgetting and adapts to existing models, though it appears incremental as it builds on Transformer-based methods.
The paper tackles the problem of enhancing text representation by incorporating large-scale out-of-domain commonsense, which faces challenges due to domain discrepancy and catastrophic forgetting. It proposes OK-Transformer, a method that integrates commonsense descriptions into text representation without requiring pre-training, achieving improvements in tasks like commonsense reasoning and text classification.
We study how to enhance text representation via textual commonsense. We point out that commonsense has the nature of domain discrepancy. Namely, commonsense has different data formats and is domain-independent from the downstream task. This nature brings challenges to introducing commonsense in general text understanding tasks. A typical method of introducing textual knowledge is continuing pre-training over the commonsense corpus. However, it will cause catastrophic forgetting to the downstream task due to the domain discrepancy. In addition, previous methods of directly using textual descriptions as extra input information cannot apply to large-scale commonsense. In this paper, we propose to use large-scale out-of-domain commonsense to enhance text representation. In order to effectively incorporate the commonsense, we proposed OK-Transformer (\underline{O}ut-of-domain \underline{K}nowledge enhanced \underline{Transformer}). OK-Transformer effectively integrates commonsense descriptions and enhances them to the target text representation. In addition, OK-Transformer can adapt to the Transformer-based language models (e.g. BERT, RoBERTa) for free, without pre-training on large-scale unsupervised corpora. We have verified the effectiveness of OK-Transformer in multiple applications such as commonsense reasoning, general text classification, and low-resource commonsense settings.