Predicting population neural activity in the Algonauts challenge using end-to-end trained Siamese networks and group convolutions
This work addresses the problem of modeling brain representations for neuroscience researchers, but it is incremental as it applies existing methods to a specific challenge.
The authors tackled the Algonauts challenge of predicting neural activity distances (RDMs) in visual brain regions by developing a customized deep learning model using Siamese networks and group convolutions, achieving best results with distances computed from the last layer.
The Algonauts challenge is about predicting the object representations in the form of Representational Dissimilarity Matrices (RDMS) derived from visual brain regions. We used a customized deep learning model using the concept of Siamese networks and group convolutions to predict neural distances corresponding to a pair of images. Training data was best explained by distances computed over the last layer.