Common Sense Beyond English: Evaluating and Improving Multilingual Language Models for Commonsense Reasoning
This work addresses the problem of limited commonsense reasoning capabilities in multilingual language models for non-English speakers, representing an incremental advancement in the field.
The authors tackled the lack of commonsense reasoning evaluation beyond English by creating multilingual datasets and a probing task, and they improved performance using multilingual contrastive pre-training, achieving significant gains on benchmarks.
Commonsense reasoning research has so far been limited to English. We aim to evaluate and improve popular multilingual language models (ML-LMs) to help advance commonsense reasoning (CSR) beyond English. We collect the Mickey Corpus, consisting of 561k sentences in 11 different languages, which can be used for analyzing and improving ML-LMs. We propose Mickey Probe, a language-agnostic probing task for fairly evaluating the common sense of popular ML-LMs across different languages. In addition, we also create two new datasets, X-CSQA and X-CODAH, by translating their English versions to 15 other languages, so that we can evaluate popular ML-LMs for cross-lingual commonsense reasoning. To improve the performance beyond English, we propose a simple yet effective method -- multilingual contrastive pre-training (MCP). It significantly enhances sentence representations, yielding a large performance gain on both benchmarks.