Scientists have developed a new synaptic transistor that works like the human brain
Scientists have made a significant breakthrough in the field of AI. Researchers from Northwestern University, Boston College and MIT have developed a new synaptic transistor that works just like the human brain.
This advanced device, capable of both processing and simultaneously storing information, marks a marked shift from traditional machine learning tasks to performing associative learning similar to higher-level human cognition, Study Finds reports.
This development represents a device that works effectively at room temperature. This is a marked improvement over previous brain-like computing devices that required extremely cold conditions to keep their circuits from overheating.
Due to its fast operation, low power consumption and the ability to store information without power, the new transistor is well suited for real-world applications.
“The brain has a fundamentally different architecture than a digital computer. In a digital computer, data is moved back and forth between the microprocessor and memory, which consumes a lot of power and creates a bottleneck when trying to multitask. On the other hand, in the brain, memory and information processing are located together and are fully integrated, which leads to an increase in energy efficiency by orders of magnitude. Our synaptic transistor similarly provides simultaneous memory and information processing to be more precise mimic the brain” – said the co-author of the study, Mark Gersam, a professor at the School of Engineering at Northwestern University.
The development of this device is taking place at a crucial time. As smart devices collect ever-increasing amounts of data, the need for efficient processing methods that don’t overload power grids is becoming ever more pressing.
Traditional digital systems that separate processing and storage are not energy efficient for processing large data sets. Although memristors (memory resistors) are the leading technology for combining processing and memory functions, they still involve power-intensive switching.
Mark Gersam and his team applied a new strategy that includes moiré patterns. It is a type of geometric design that is created when two patterns are superimposed on each other.
By stacking two-dimensional materials such as bilayer graphene and hexagonal boron nitride and twisting them to form a moiré pattern, the researchers were able to manipulate the electronic properties of the graphene layers. This manipulation made it possible to create a synaptic transistor with enhanced neuromorphic functionality at room temperature.
“If artificial intelligence is designed to imitate human thought, then one of the lowest-level tasks will be to classify data, that is, simply sort it into containers. Our goal is to advance artificial intelligence technology in the direction of higher level thinking. Real-world conditions are often more complex than current artificial algorithms can handle intelligence Therefore, we tested our new devices in more challenging conditions to verify their enhanced capabilities.” Gersam explained.
Testing the device included learning to recognize patterns and similarities (a form of associative learning).
For example, if you teach it to identify a pattern like “000,” the transistor will be able to determine that “111” is more like “000” than “101,” demonstrating a higher level of cognitive function. This ability to process complex and imperfect input data is of great importance for real-world AI applications, such as improving the reliability of self-driving vehicles in challenging environments.
The research demonstrates a paradigm shift in electronics, especially with regard to AI and machine learning tasks. By moving away from traditional silicon architecture and exploring the physics of moiré patterns, experts have opened up a new realm of computing hardware, paving the way for more sophisticated and energy-efficient AI technologies.
As a reminder, recently a team of researchers from Brown University and several Chinese universities conducted an experiment to find out whether artificial intelligence bots working on the basis of the ChatGPT 3.5 model can complete the software development process without prior training.
Subscribe to ProIT on Telegram so you don’t miss a post!