How aligned are different alignment metrics?

Published: 02 Mar 2024, Last Modified: 10 May 2024ICLR 2024 Workshop Re-Align PosterEveryoneRevisionsBibTeXCC BY 4.0
Track: short paper (up to 5 pages)
Keywords: alignment, brain-score, integrative benchmarking
TL;DR: We investigate how aligned different brain-machine alignment metrics are and investigate new ways of aggregating such metrics.
Abstract: In recent years, various methods and benchmarks have been proposed to empirically evaluate the alignment of artificial neural networks to human neural and behavioral data. But how aligned are different alignment metrics? To answer this question, we here analyze visual data from Brain-Score (Schrimpf et al., 2018), including metrics from the model-vs-human toolbox (Geirhos et al., 2021), together with human feature alignment (Linsley et al., 2018; Fel et al., 2022) and human similarity judgements (Muttenthaler et al., 2022). We find that pairwise correlations between neural scores and behavioral scores are quite low and sometimes even negative. For instance, the average correlation between those 80 models on Brain-Score that were fully evaluated on all 69 alignment metrics we considered is only 0.198. Assuming that all of the employed metrics are sound, this implies that alignment with human perception may best be thought of as a multidimensional concept, with different methods measuring fundamentally different aspects. Our results underline the importance of integrative benchmarking, but also raise questions about how to correctly combine and aggregate individual metrics. Aggregating by taking the arithmetic average, as done in Brain-Score, leads to the overall performance currently being dominated by behavior (95.25% explained variance) while the neural predictivity plays a less important role (only 33.33% explained variance). As a first step towards making sure that different alignment metrics all contribute fairly towards an integrative benchmark score, we therefore conclude by comparing three different aggregation options.
Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.
Submission Number: 81