Scientists have developed a groundbreaking new transistor design that mimics the architecture of the human brain, enabling the storage and processing of information in a way that most artificial intelligence (AI) systems cannot achieve. This new technology, known as “synaptic transistor,” integrates computing power and memory, eliminating the need for separate components as in traditional computer architecture. The synaptic transistor not only achieves significantly higher energy efficiency but also allows for extremely fast data flow.
Previously, synaptic transistors based on classical electronics and using silicon substrates were used for construction. However, these transistors could only operate at extremely low temperatures. The new design overcomes this limitation by utilizing bilayer graphene (BLG) and hexagonal boron nitride (hBN), set at a specific twist angle to create a moiré pattern. This pattern creates new electronic properties that do not exist in any individual layer. With a near-perfect match between hBN and BLG, researchers were able to achieve functionality at room temperature.
The researchers trained the system by exposing it to data in order to learn to recognize patterns. Then, new sequences were shown to it that were similar to the training data but not identical, allowing it to perform associative learning. This is a cognitive task that most machine learning systems struggle with. The system successfully detected patterns and identified sequences based on the associations it had learned, even when presented with incomplete patterns.
“Our goal is to develop AI technology towards higher-level thinking,” said Mark Hersam, co-lead researcher and professor of materials science, engineering, and computer science at Northwestern University. With the potential to create highly energy-efficient circuits, this new transistor design could revolutionize artificial intelligence systems and machine learning in the future, enabling them to handle more complex real-world conditions and perform advanced cognitive tasks.
Source: Adapted from the original article on domain.com.
The source of the article is from the blog reporterosdelsur.com.mx