From Frege to chatGPT: Compositionality in language, cognition, and deep neural networks

J Russin, SW McGrath, DJ Williams… - arXiv preprint arXiv …, 2024 - arxiv.org
Compositionality has long been considered a key explanatory property underlying human
intelligence: arbitrary concepts can be composed into novel complex combinations …

Toward Compositional Behavior in Neural Models: A Survey of Current Views

K McCurdy, P Soulos, P Smolensky… - Proceedings of the …, 2024 - aclanthology.org
Compositionality is a core property of natural language, and compositional behavior (CB) is
a crucial goal for modern NLP systems. The research literature, however, includes …

Mechanisms of Symbol Processing for In-Context Learning in Transformer Networks

P Smolensky, R Fernandez, ZH Zhou, M Opper… - arXiv preprint arXiv …, 2024 - arxiv.org
Large Language Models (LLMs) have demonstrated impressive abilities in symbol
processing through in-context learning (ICL). This success flies in the face of decades of …

Terminating Differentiable Tree Experts

J Thomm, M Hersche, G Camposampiero… - … Conference on Neural …, 2024 - Springer
We advance the recently proposed neuro-symbolic Differentiable Tree Machine, which
learns tree operations using a combination of transformers and Tensor Product …

Special Session: Neuro-Symbolic Architecture Meets Large Language Models: A Memory-Centric Perspective

M Ibrahim, Z Wan, H Li, P Panda… - 2024 International …, 2024 - ieeexplore.ieee.org
Large language models (LLMs) have significantly transformed the landscape of artificial
intelligence, demonstrating exceptional capabilities in natural language understanding and …

Discrete Dictionary-based Decomposition Layer for Structured Representation Learning

T Park, HC Kim, M Lee - arXiv preprint arXiv:2406.06976, 2024 - arxiv.org
Neuro-symbolic neural networks have been extensively studied to integrate symbolic
operations with neural networks, thereby improving systematic generalization. Specifically …

Rules, frequency, and predictability in morphological generalization: behavioral and computational evidence from the German plural system

K McCurdy, K McCurdy - 2024 - era.ed.ac.uk
Morphological generalization, or the task of mapping an unknown word (such as a novel
noun Raun) to an inflected form (such as the plural Rauns), has historically proven a …

Recurrent Transformers Trade-off Parallelism for Length Generalization on Regular Languages

P Soulos, A Terzic, M Hersche… - The First Workshop on …, 2024 - openreview.net
Transformers have achieved remarkable success in Natural Language Processing but
struggle with state tracking and algorithmic reasoning tasks, such as modeling Regular …