Neural Networks for Abstraction & Reasoning
Mikel Bober-Irizar · Soumya Banerjee
2024 Poster
in
Workshop: Multimodal Algorithmic Reasoning Workshop
in
Workshop: Multimodal Algorithmic Reasoning Workshop
Abstract
For decades, AI research has aimed to replicate human-like abstraction and reasoning, yet broad generalization beyond training distributions remains elusive. The Abstraction & Reasoning Corpus (ARC), a dataset designed to test broad generalisation, has remained unsolved after five years, with the best solver based on handcrafted rules. We adapt two novel approaches based on neural networks: a neurosymbolic reasoning system based on DreamCoder, and a series of frontier large-language models, and compare their performance.
Chat is not available.
Successful Page Load