Semi-supervised sequence modeling for improved behavioral segmentation
M. R. Whiteway, E. S. Schaffer, A. Wu, E. K. Buchanan, O. F. Onder, N. Mishra, L. Paninski
This paper introduces a semi-supervised sequence model for behavioral segmentation that combines supervised, weakly supervised, and self-supervised losses to improve accuracy with sparse labels.
bioRxiv · Jun 16, 2021 · PreprintAnimal Behavior & Cognition