Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2401.16024 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914656483803136 |
|---|---|
| author | Hersche, Michael di Stefano, Francesco Hofmann, Thomas Sebastian, Abu Rahimi, Abbas |
| author_facet | Hersche, Michael di Stefano, Francesco Hofmann, Thomas Sebastian, Abu Rahimi, Abbas |
| contents | Abstract reasoning is a cornerstone of human intelligence, and replicating it with artificial intelligence (AI) presents an ongoing challenge. This study focuses on efficiently solving Raven's progressive matrices (RPM), a visual test for assessing abstract reasoning abilities, by using distributed computation and operators provided by vector-symbolic architectures (VSA). Instead of hard-coding the rule formulations associated with RPMs, our approach can learn the VSA rule formulations (hence the name Learn-VRF) with just one pass through the training data. Yet, our approach, with compact parameters, remains transparent and interpretable. Learn-VRF yields accurate predictions on I-RAVEN's in-distribution data, and exhibits strong out-of-distribution capabilities concerning unseen attribute-rule pairs, significantly outperforming pure connectionist baselines including large language models. Our code is available at https://github.com/IBM/learn-vector-symbolic-architectures-rule-formulations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_16024 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Probabilistic Abduction for Visual Abstract Reasoning via Learning Rules in Vector-symbolic Architectures Hersche, Michael di Stefano, Francesco Hofmann, Thomas Sebastian, Abu Rahimi, Abbas Machine Learning Artificial Intelligence Abstract reasoning is a cornerstone of human intelligence, and replicating it with artificial intelligence (AI) presents an ongoing challenge. This study focuses on efficiently solving Raven's progressive matrices (RPM), a visual test for assessing abstract reasoning abilities, by using distributed computation and operators provided by vector-symbolic architectures (VSA). Instead of hard-coding the rule formulations associated with RPMs, our approach can learn the VSA rule formulations (hence the name Learn-VRF) with just one pass through the training data. Yet, our approach, with compact parameters, remains transparent and interpretable. Learn-VRF yields accurate predictions on I-RAVEN's in-distribution data, and exhibits strong out-of-distribution capabilities concerning unseen attribute-rule pairs, significantly outperforming pure connectionist baselines including large language models. Our code is available at https://github.com/IBM/learn-vector-symbolic-architectures-rule-formulations. |
| title | Probabilistic Abduction for Visual Abstract Reasoning via Learning Rules in Vector-symbolic Architectures |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2401.16024 |