Predicting Human Similarity Judgments Using Large Language Models
This provides a more efficient method for researchers in psychology, neuroscience, and machine learning to access mental representations, though it is incremental as it builds on existing approximation procedures.
The paper tackled the problem of expensive human similarity judgment collection by predicting them using text descriptions from large language models, showing that their models outperform previous visual-based approaches on six naturalistic image datasets.
Similarity judgments provide a well-established method for accessing mental representations, with applications in psychology, neuroscience and machine learning. However, collecting similarity judgments can be prohibitively expensive for naturalistic datasets as the number of comparisons grows quadratically in the number of stimuli. One way to tackle this problem is to construct approximation procedures that rely on more accessible proxies for predicting similarity. Here we leverage recent advances in language models and online recruitment, proposing an efficient domain-general procedure for predicting human similarity judgments based on text descriptions. Intuitively, similar stimuli are likely to evoke similar descriptions, allowing us to use description similarity to predict pairwise similarity judgments. Crucially, the number of descriptions required grows only linearly with the number of stimuli, drastically reducing the amount of data required. We test this procedure on six datasets of naturalistic images and show that our models outperform previous approaches based on visual information.