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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Despite the progress in deep learning networks, efficient learning at theedge (enabling adaptable, low-complexity machine learning solutions) remains acritical need for defense and commercial applications. We envision a pipelineto utilize large neuroimaging datasets, including maps of the brain whichcapture neuron and synapse connectivity, to improve machine learningapproaches. We have pursued different approaches within this pipelinestructure. First, as a demonstration of data-driven discovery, the team hasdeveloped a technique for discovery of repeated subcircuits, or motifs. Thesewere incorporated into a neural architecture search approach to evolve networkarchitectures. Second, we have conducted analysis of the heading directioncircuit in the fruit fly, which performs fusion of visual and angular velocityfeatures, to explore augmenting existing computational models with new insight.Our team discovered a novel pattern of connectivity, implemented a new model,and demonstrated sensor fusion on a robotic platform. Third, the team analyzedcircuitry for memory formation in the fruit fly connectome, enabling the designof a novel generative replay approach. Finally, the team has begun analysis ofconnectivity in mammalian cortex to explore potential improvements totransformer networks. These constraints increased network robustness on themost challenging examples in the CIFAR-10-C computer vision robustnessbenchmark task, while reducing learnable attention parameters by over an orderof magnitude. Taken together, these results demonstrate multiple potentialapproaches to utilize insight from neural systems for developing robust andefficient machine learning techniques.

Exploiting Large Neuroimaging Datasets to Create Connectome-Constrained Approaches for more Robust, Efficient, and Adaptable Artificial Intelligence