Deploying HD-RNN

The most unique aspect of Binatix's deep learning technology is its approach to robust modeling of complex data. Mimicking the efficiency and robustness by which the human brain analyzes and represents information has been a core challenge in artificial intelligence (AI) research for decades. For instance, humans are exposed to massive amounts of visual and auditory data every second of every day, and are able to capture critical aspects of it in a way that allows for appropriate feature recollection and action selection. For decades, it has been known that the brain is a massively parallel fabric, in which computation processes and memory storage are highly distributed.

HD-RNN architecture naturally lends itself to massive parallelism which bodes well with commercial availability of affordable GPU (graphical processing unit) add-on cards that plug into standard PC hardware.

A server can be equipped with multiple GPU cards, resulting in compute capacity of thousands of processors. Binatix provides an ultra-fast implementation of HD-RNN on such cards (currently NVidia GPUs) where compute-intense training of the deep learning framework can be done in hours instead of weeks. Also, once HD-RNN is trained, it can be deployed on small footprint hardware that fit the intended application.