Reproduce “Modelling Analogies and Analogical Reasoning: Connecting Cognitive Science Theory and NLP Research” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Molly R Petersen, Claire E Stevenson, Lonneke van der Plas
This paper integrates cognitive science theories of human analogical reasoning with natural language processing research, offering a critical framework to evaluate and guide AI models toward deeper relational understanding. It identifies conceptual parallels between human cognitive processes and NLP approaches while highlighting current models' limitations in complex, human-like analogy tasks.
Key insights
- Core cognitive processes of analogical reasoning have direct conceptual counterparts in NLP, but current computational approaches often lack deep theoretical grounding from cognitive science.
- NLP models demonstrate strong performance on simple morpho-syntactic analogies but struggle with more complex symbolic, semantic, and contextual analogy types that demand deeper relational understanding.
- Bridging cognitive science and NLP can provide a theoretical framework to address fundamental AI limitations, fostering the development of more robust, interpretable, and generalizable models focused on relational intelligence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.