AI Insight
Researchers have developed Similarity-Based Representation Factorization (SRF), a computational method that extracts interpretable, low-dimensional patterns from similarity matrices in representational data. The method was validated across simulations and multiple datasets from neuroscience, psychology, and artificial intelligence, demonstrating it can recover meaningful dimensions even from incomplete or sparsely sampled data. SRF outperformed traditional similarity matrix comparison methods in hypothesis testing and matched results from task-specific models while also predicting independent behavioral properties.
Why it matters
This method provides a unified approach for analyzing how the brain, behavior, and artificial systems represent information, making it easier for researchers across disciplines to interpret complex data patterns. The ability to work with incomplete data and extract interpretable dimensions could accelerate discovery in neuroscience and improve our understanding of both biological and artificial intelligence systems.
arXiv:2605.26921v2 Announce Type: replace-cross
Abstract: The study of representations is widespread across fields, including neuroscience, psychology, and artificial intelligence. While representations are often studied and compared through similarities between stimuli, current methods provide only limited access to the dimensions that shape these representations and are often limited in interpretability. To overcome these challenges, here we introduce Similarity-Based Representation Factorization (SRF), a general computational method for recovering low-dimensional, non-negative, interpretable embeddings from similarity matrices derived from measured data. Across simulations and many neural, behavioral, and computational datasets, SRF recovers interpretable dimensions from diverse forms of representational data, even for very sparsely sampled, incomplete data. The dimensions derived from these datasets match those obtained by task-specific models, predict independent behavioral properties, improve exploratory analysis, and offer higher power for confirmatory hypothesis testing than comparing similarity matrices. Together, these results establish SRF as a general-purpose method with broad applications for uncovering, understanding, and using the dimensions underlying representations.