High-Dimensional Computing, SDM Memory & Hyperscaling - A session on the Cloud’s missing AI component
It has become strikingly evident that memory is one of the most prominent elements of human cognition and one that has become increasingly important for artificial intelligence (AI) systems. As AI agents evolve into the third wave and tackle more complex scenarios, the role of memory is not only relevant but has become mandatory. Paradoxically, memory models are one of the most notably missing components of AI platforms and frameworks. The session provides a brief review of conventional memory use in today’s server based systems followed by an introduction and description of an architecture based on High-Dimensional SDM memory called the “Kanerva Partition” and how HD Computing will transform future Machine Learning and Artificial General Intelligence systems.
• Near Data Processing Acceleration
• Predictive Analytics
• Neuromorphic Architectures
• In-Memory Computing Cognitive Extensions
Current research is focused on investigating the adaption of In-Memory Compute machine server architecture to High-Dimensional Computing within a standard rack model space – leading to the development of a standard hyperscale cognitive base element for Cloud Computing.