European Commission logo
polski polski
CORDIS - Wyniki badań wspieranych przez UE
CORDIS

A neural network builder with remotely controlled parallel computing

Periodic Reporting for period 1 - NEUROCLOUD (A neural network builder with remotely controlled parallel computing)

Okres sprawozdawczy: 2017-09-01 do 2019-02-28

Computer simulations of neural networks have been a driving force behind our advances in artificial intelligence but but also a critical instrument in our understanding of the human brain. Building and exploring such networks usually engages substantial computational resources requiring specific programing skills and dedicated facilities. Aiming to facilitate high-end neural network exploration among scientific and industrial users, we have developed novel computational environment ARACHNE. ARACHNE enables an investigator to construct and explore cellular networks of arbitrary biophysical and architectural complexity using a user-friendly interface on a local computer or a mobile device. Through small-file exchange via the internet, the interface controls computations on a remote computer cluster or cloud. Importantly, ARACHNE can integrate traditional wired network connections with a diffuse type of inter-cellular signalling: the latter reflects the emerging role of glial cells in brain processing. In the present project we have validated the usability, efficiency, and reloability of ARACHNE design in the open-field user environment, and created a self-extracting package for its download and installation by untrained users.

We have determined the market potential and the commercial viability of ARACHNE and established relevant industry-standard provisions. As a result, we have created a private company (Neuroalgebra Limited) and its initial website facilities (neuroalgebra.com) for the commercial exploitation of ARACHNE and associated software products. The company provides a full version of ARACHNE free of charge, to comply with licensing regulations, while providing chargeable on-line facilities to run the package on a computer cloud by IT-untrained users.

The economical impact of the project is two-fold. Firstly, ARACHNE provides an advanced tool for academic, medical or other exploratory or industrial research groups that rely on high-end neural network simulations but have limited capacities for on-site large-scale computer architecture and designated IT staff. Secondly, the newly developed software architecture that enables ARACHNE to control cloud computer simulations from a mobile console via the internet (or mobile operators), could potentially provide a patentable IT technology for a wide range of large-scale computing applications.

The societal impact of the project is that ARACHNE provides a novel and unique tool to boost our understanding of the human brain. In addition to having a clear impact on the general knowledge, this quest is important for advancing novel therapeutic strategies to deal with common neurological conditions, such as epilepsy or dementia. In addition, by facilitating the exploratory quest of complex neural networks, ARACHNE boosts the field of artificial intelligence, which is a key area of human endeavour. Finally, the ARACHNE architecture which enables on-line or mobile control of cloud computing should facilitate and widen public access to high-end computations.