Can language models learn analogical reasoning? Investigating training objectives and comparisons to human performance
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909188302569472 |
|---|---|
| author | Petersen, Molly R. van der Plas, Lonneke |
| author_facet | Petersen, Molly R. van der Plas, Lonneke |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_05597 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Can language models learn analogical reasoning? Investigating training objectives and comparisons to human performance Petersen, Molly R. van der Plas, Lonneke Computation and Language 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. |
| title | Can language models learn analogical reasoning? Investigating training objectives and comparisons to human performance |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2310.05597 |