A neural network which bridges the computation, learning, and application’s implementation gap across conventional neural networks and cognitive learning processes with dual-band training and application layers.
AFNN has application where quick learning is paramount while some precision can be sacrificed. Error reduces rapidly and achieves conventional error bandwidth with increasing number of iterations.
AFNN Engine titled MyND wraps AFNN into an ML tool that integrates AFNN into real life applications. MyND is currently being successfully applied in Spectrix game streaming application, Mossaik XDR photo editor and the graphics processing Freznel engine.
ISSN electronic - 1876-1119
ISBN - 978-981-13-6771-7
ISBN eBook - 978-981-13-6772-4
An advanced graphics processing engine that uses top of the technological developments and provides Extreme Dynamic Range image rendering, supports 64 bit image processing and optimizes graphics compute load balancing between CPU, Integrated GPU or the dedicated GPU.
Freznel engine is being used in development of Mossaik XDR photo editor.
Utilizing the best of H.265 encoding & compression, Mobile telephony (5G) and GPU processing, Sapphire Streamworks delivers extremely low latency gaming experience or high end graphics manipulation on wirelessly connected remote devices. The technology is being developed to bring the games that require complex computations to a handheld mobile platform.
PixStream Engine - Applying the principles of AFNN, software adapts to the dynamic data bit rate and delivers seamless user experience with regards to latency by optimizing the resolution and reducing data requirements where background image is not changing relative to the other parts in the image.