AIBrainWideBench announced to evaluate neural-activity models across animals and tasks
A project co-lead calls rigorous, standardized evaluation a major bottleneck in modeling neural data, arguing that techniques need testing under multiple regimes and downstream tasks.
TLDR
A project co-lead announced BrainWideBench, a new benchmark for evaluating how models of neural activity generalize across animals and tasks. The announcement presents standardized evaluation as a major challenge for the field and links to the team's paper on arXiv.
BrainWideBench announced to evaluate neural-activity models across animals and tasks
A project co-lead calls rigorous, standardized evaluation a major bottleneck in modeling neural data, arguing that techniques need testing under multiple regimes and downstream tasks.