Join Us 
Revolutionizing AI: A groundbreaking study led by researchers from UCL and Imperial College London is propelling us toward a new era of energy-efficient computing inspired by the intricacies of the human brain. In their pursuit of sustainable computing solutions, the team utilized chiral (twisted) magnets as the foundation for computational processes. Published in the prestigious journal Nature Materials, the research introduces the concept of physical reservoir computing, a paradigm that leverages the inherent physical properties of materials to drastically reduce energy consumption.
The crux of the study lies in the adaptability of these chiral magnets to various machine-learning tasks. By applying an external magnetic field and manipulating temperature, the researchers successfully tailored the physical properties of the materials, overcoming the previous limitations associated with physical reservoir computing. This breakthrough could revolutionize the field, offering not only energy-efficient computing but also adaptability across a spectrum of tasks, akin to the multifaceted nature of human brain functions.
Dr. Oscar Lee, the lead author of the study, emphasizes the significance of this work in realizing the full potential of physical reservoirs. He envisions a future where computers not only demand significantly less energy but also possess the ability to optimize their computational properties for different tasks, mimicking the versatility of the human brain.
Traditional computing models, characterized by separate units for data storage and processing, consume substantial electricity, resulting in inefficiencies and excess heat production. Physical reservoir computing, a neuromorphic approach inspired by the brain’s functioning, seeks to eliminate the need for distinct memory and processing units, offering a more sustainable and efficient means of data processing.
The study investigated the energy absorption of chiral magnets in different magnetic phases, unveiling their proficiency in various computing tasks. The skyrmion phase exhibited potent memory capacity suitable for forecasting tasks, while the conical phase, with its non-linearity, proved ideal for transformation tasks and classification.
Co-author Dr. Jack Gartside of Imperial College London highlights the collaborative effort to identify promising materials for unconventional computing, emphasizing the intricate magnetic textures that make these materials unique. The study’s innovative approach demonstrates how reconfiguring physical phases can directly tailor the performance of neuromorphic computing.
As the next step, researchers aim to identify commercially viable and scalable materials and device architectures, paving the way for a new era of energy-efficient and adaptable computing solutions. The study, conducted in collaboration with researchers in Japan and Germany, received support from various entities, including the Leverhulme Trust, Engineering and Physical Sciences Research Council (EPSRC), Imperial College London President’s Excellence Fund for Frontier Research, Royal Academy of Engineering, Japan Science and Technology Agency, Katsu Research Encouragement Award, Asahi Glass Foundation, and the German Research Foundation (DFG).
[tta_listen_btn]
