Can language models learn analogical reasoning? Investigating training objectives and comparisons to human performance
This addresses the problem of evaluating and enhancing analogical reasoning in AI systems, which is incremental as it builds on existing NLP benchmarks.
The paper investigates whether language models can learn analogical reasoning by testing various training objectives on human-like analogies, finding that models achieve this with minimal data and approach human performance levels.
While analogies are a common way to evaluate word embeddings in NLP, it is also of interest to investigate whether or not analogical reasoning is a task in itself that can be learned. In this paper, we test several ways to learn basic analogical reasoning, specifically focusing on analogies that are more typical of what is used to evaluate analogical reasoning in humans than those in commonly used NLP benchmarks. Our experiments find that models are able to learn analogical reasoning, even with a small amount of data. We additionally compare our models to a dataset with a human baseline, and find that after training, models approach human performance.