English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We present an exploration of machine learning architectures for predictingbrain responses to realistic images on occasion of the Algonauts Challenge2023. Our research involved extensive experimentation with various pretrainedmodels. Initially, we employed simpler models to predict brain activity butgradually introduced more complex architectures utilizing available data andembeddings generated by large-scale pre-trained models. We encountered typicaldifficulties related to machine learning problems, e.g. regularization andoverfitting, as well as issues specific to the challenge, such as difficulty incombining multiple input encodings, as well as the high dimensionality, unclearstructure, and noisy nature of the output. To overcome these issues we testedsingle edge 3D position-based, multi-region of interest (ROI) and hemispherepredictor models, but we found that employing multiple simple models, eachdedicated to a ROI in each hemisphere of the brain of each subject, yielded thebest results - a single fully connected linear layer with image embeddingsgenerated by CLIP as input. While we surpassed the challenge baseline, ourresults fell short of establishing a robust association with the data.

Exploration and Comparison of Deep Learning Architectures to Predict Brain Response to Realistic Pictures