The research team developed a new method for pattern recognition called “reservoir computing.” They used a vortex of magnons to enable near instantaneous algorithmic functions.
The researchers not only developed and tested this new reservoir material, but also demonstrated the potential for neuromorphic computing to work on a standard CMOS chip. This could transform blockchain and artificial intelligence (AI).
Classical computers, like those in smartphones and laptops, use binary transistors that are either on or off (1 or 0). Neuromorphic computers instead use programmable artificial neurons to mimic organic brain activity. They send signals across varying neuron patterns over time, not just 1s and 0s.
This matters for blockchain and AI because neuromorphic computers are well-suited for pattern recognition and machine learning algorithms. Binary systems use boolean algebra for computations. So classical computers remain superior for crunching numbers. However, they struggle with pattern recognition, especially with incomplete or noisy data.
Classical systems require considerable time to solve complex cryptographic puzzles. They cannot solve problems where insufficient data prevents a math-based solution. Sectors like finance, AI, and transportation have endless real-time data. Classical computers struggle with obscured problems that cannot be reduced to true/false computations. For example, driverless car development has proven difficult to frame as a classical computing problem due to incomplete data.
Neuromorphic computers are designed to handle problems with missing information. In transportation, a classical computer cannot predict traffic flow due to too many variables. Neuromorphic computers constantly react to real-time data since they do not process data points individually.
Instead, neuromorphic computers run data through changing pattern configurations like human brains. The main benefit is their extremely low power consumption compared to classical and quantum computing. This could significantly reduce the cost of operating blockchains and mining new blocks.
Neuromorphic computers could also provide major speed increases for machine learning, especially involving real-world sensors or real-time data processing.
#Blockchain #AI #Supercomputer
