Computers have advanced significantly in terms of their power and capabilities, surpassing human brains in certain tasks but still lagging behind in terms of energy efficiency. The energy requirements of computers are much higher compared to the human brain for tasks such as image processing and recognition. With the increasing global energy consumption by on-chip electronics, the need for more energy-efficient computing technologies is becoming urgent, especially in the context of global warming.
Neuromorphic computing has emerged as a promising solution to bridge the energy efficiency gap. By mimicking the structure and operations of the human brain, where processing occurs in parallel across low power-consuming neurons, it may be possible to achieve brain-like energy efficiency. Researchers from UC Santa Barbara and Intel Labs have proposed an ultra-energy efficient platform using 2D transition metal dichalcogenide-based tunnel-field-effect transistors (TFETs) that could bring the energy requirements closer to that of the human brain.
The concept of neuromorphic computing has been around for decades, but recent advances in circuitry and transistor technologies have enabled the development of more energy-efficient and brain-inspired computing platforms. The researchers’ 2D tunnel-transistors are responsive at low voltages and mimic the energy efficient operations of the human brain. These transistors have lower off-state currents and a low subthreshold swing, allowing for faster and more efficient switching operations.
Neuromorphic computing architectures operate with sparse firing circuits, firing only when necessary, unlike conventional computers that continuously draw power throughout their operation. The energy efficiency of these systems is constrained by leakage currents when transistors are off. By using tunneling transistors with lower off-state currents, the researchers have demonstrated improved power efficiency in neuromorphic circuits, making them a promising candidate for the next generation of brain-inspired computing.
The TFETs integrated into neuromorphic circuits have shown to be more energy efficient than state-of-the-art MOSFETs. While TFETs are still in the experimental stage, their performance and energy efficiency make them a promising technology for future brain-inspired computing platforms. The researchers believe that these 2D-TFET based neuromorphic circuits could eventually be developed into three-dimensional versions to provide even closer emulation of the human brain, leading to significant improvements in energy efficiency compared to current computing technologies.
Overall, the development of energy-efficient computing technologies, such as neuromorphic computing based on 2D-TFETs, holds great promise for reducing energy consumption in chips and moving towards brain-like energy efficiency. With ongoing research and advancements in transistor technologies, the vision of highly energy-efficient and brain-inspired computing platforms may soon become a reality, paving the way for a more sustainable and efficient future.