Building computational models for the human perception system is an important step towards the ultimate goal of better understanding human intelligence and guiding artificial intelligence engineering.
Neural responses in the visual cortex were recorded using functional neuroimaging when participants were watching naturalistic videos. (Algonauts Challenge 2021)
We develop an deep learning model that combines feature representations from multiple perspectives and modalities, including image streams, motion, edges and audio features. We show that representations from each perspective separately improves the prediction performance of the final encoding model.
We are also working on prediction and intervention with brain control theory, more details will be shared.