The most popular and comprehensive Open Source ECM platform
Let us know more
Neuromorphic Chips: Technology that Emulates the Brain
Neuromorphic chips are designed to mimic the neural processing of the human brain. The technology that is being used to build them come from a variety of fields, including biology, mathematics, physics, computer science and electrical engineering.
Applications include the logistics, food automation, agriculture, and the continuous monitoring of industrial machines.
Current computer chips work by sending data back and forth between the central processor and memory chips, a computing model known as the von Neumann architecture. Neuromorphic chips are modeled around connections of a neural network model and stimuli that target different connections.
As obstacles for further improvements to standard von Neumann processing mount, neuromorphic technology has characteristics of an attractive alternative. Standard chips have steadily improved by shrinking sizes and increasing clock rates, but continuing to do that is becoming difficult and costly as the technology is running into physical limitations. Data sizes for applications like AI and Big Data are growing, putting stress on memory, and faster clock rates is also leading to problems with cooling.
Research into neuromorphic computing is speeding up. The neuromorphic chip market is about $20 million in 2019 and will reach $610 million by 2026, according to estimates by Market Report Neuromorphic chips are now available from companies like Intel, IBM and Brainchip.
Pierre Cambou, Principal Analyst for Imaging at Yole, said that “while deep learning needs huge data sets, neuromorphic learns extremely quickly from only a few images or a few words and understands time… We live in a world of interactions, and neuromorphic will be very strong in giving computers the understanding of unstructured environments.”
