The US military's research wing, DARPA, is seeking help with a new project that would allow computers to imitate the human neocortex, the region of the brain we use for language, reasoning, and perception. What could go wrong?
Though we already have machine learning algorithms that allow computers to make decisions based on constantly-shifting data, DARPA wants more. They hope for a "cortical computation model" that will allow computers to recognize new data when it's relevant to their tasks, and to adapt based on new inputs from their environments. Though cortical processors sound like something out of a 1980s scifi movie, they're exactly what a robot or drone in the field will need to gather intel and make decisions based on what it learns.
Although a thorough understanding of how the cortex works is beyond current state of the art, we are at a point where some basic algorithmic principles are being identified and merged into machine learning and neural network techniques. Algorithms inspired by neural models, in particular neocortex, can recognize complex spatial and temporal patterns and can adapt to changing environments. Consequently, these algorithms are a promising approach to data stream filtering and processing and have the potential for providing new levels of performance and capabilities for a range of data recognition problems. The cortical computational model should be fault tolerant to gaps in data, massively parallel, extremely power efficient, and highly scalable. It should also have minimal arithmetic precision requirements, and allow ultra-dense, low power implementations.
Read more via Network World