Physical reservoir computing can be used to perform high-speed processing for artificial intelligence with low power consumption.
Researchers from Japan design a physically tunable reservoir device based on dielectric relaxation at the ion-electrode-liquid interface.
In the near future, more and more artificial intelligence processing will need to take place at the edge – closer to the user and where data is collected, rather than on a remote computer server. This will require high speed data processing with low power consumption. Physical reservoir computing is an intriguing platform for this purpose, and a new breakthrough from scientists in Japan has just made this much more versatile and practical.
Physical reservoir computing (PRC), based on the transient response of physical systems, is an attractive machine learning framework that can perform high-speed processing of power-level time series signals. short. However, the PRC system has a low test capacity, limiting the signals it can handle. Now, researchers from Japan present ionic liquids as an easily tunable physical container that can be optimized for signal processing over a wide range of time by varying their viscosity.
Artificial intelligence (AI) is rapidly gaining popularity in modern society and will be deployed more widely in the coming years. In applications involving sensors and internet of things devices, the standard is often advanced AI, a technology in which computation and analysis are performed close to the user (where the data is collected). and not far on a centralized server. This is because edge AI has low power requirements as well as high-speed data processing capabilities, characteristics that are particularly desirable in real-time processing of time series data.
Time scale of signals typically generated in a live environment. The response time of the ionic liquid PRC system developed by the team can be tuned to optimize for processing such real-world signals. Credit: Kentaro Kinoshita from TUS
In this regard, physical reservoir computing (PRC), based on the transient dynamics of physical systems, can greatly simplify the computational models of advanced AI. This is because PRCs can be used to store and process analog signals into edges where AI can work and analyze efficiently. However, the dynamics of solid PRC systems are characterized by specific time periods that are not easily tuned and are often too fast for most physical signals. This mismatch in time scales and their low controllability renders PRCs largely unsuitable for real-time signal processing in live environments.
To tackle this problem, a research team from Japan with the participation of Professors Kentaro Kinoshita and Sang-Gyu Koh, a PhD student from Tokyo University of Science, and senior researchers Dr. Hiroyuki Akinaga, Dr. Hisashi Shima and Dr. Yasuhisa Naitoh from the National Institute of Advanced Industrial and Industrial Sciences, suggest, in a new study published in the journal Scientific reports, the use of liquid PRC system instead. Professor Kinoshita explains: “Replacing conventional solid tanks with liquid ones will lead to AI devices that can learn directly at the time scale of signals generated from the environment, such as voice and vibration”. “Ionic liquids are stable molten salts made up entirely of freely moving charges. The dielectric expansion of ionic liquids, or the way its charges rearrange in response to an electrical signal, could be used as a reservoir and hold great promise for advanced AI physics computation. up. “
The response of the ionic liquid PRC system can be tuned to be optimized for processing a wide range of signals by varying its viscosity through tuning the cationic side chain length. Credit: Kentaro Kinoshita from TUS
In their study, the team designed a PRC system with an ionic liquid (IL) of the organic salt, 1-alkyl-3-methylimidazolium bis (trifluoromethane sulfonyl) imide ([Rmim+] [TFSI–] R = ethyl(e), butyl(b), hexyl(h), and octyl(o)), whose cationic (positively charged ions) part can easily change with the length of the chosen alkyl chain. They fabricated gold gap electrodes and filled the gaps with IL. “We found that the duration of the reservoir, although complex in nature, can be directly controlled by the viscosity of the IL, which depends on the length of the alkyl cation chain. “Changing the alkyl group in organic salts is easy to implement and gives us a controllable system that can be specified for a wide range of signal life cycles, enabling a wide range of applications,” said Professor Kinoshita. computer use in the future. By tuning the alkyl chain length from 2 to 8 units, the researchers achieved characteristic response times that ranged from 1 to 20 µs, with longer alkyl chains resulting in longer response times. and the device’s adjustable AI learning performance.
The testability of the system has been demonstrated using the AI image recognition task. The AI was presented with a handwritten image as input, represented by a rectangular pulse voltage of 1 µs width. By increasing the side chain length, the team made the kinetics of the target signal temporarily approach the target signal, with improved discrimination rates for higher chain lengths. This is because, compared to [emim+] [TFSI–]where the current dilates to its value in about 1 µs, the IL with longer side chains and thus, longer elongation time retains the history of the time series data better, improving identification[{” attribute=””>accuracy. When the longest sidechain of 8 units was used, the discrimination rate reached a peak value of 90.2%.
Input signal conversion through the ionic liquid-based PRC system. The reservoir output in the form of current response (top and middle) to an input voltage pulse signal (bottom) are shown. If the current decay (dielectric relaxation) is too fast/slow, it reaches its saturation value before the next signal input and no history of the previous signal is retained (middle image). Whereas, if the current response attenuates with a relaxation time that is properly matched with the time scales of the input pulse, the history of the previous input signal is retained (top image). Credit: Kentaro Kinoshita from TUS
These findings are encouraging as they clearly show that the proposed PRC system based on the dielectric relaxation at an electrode-ionic liquid interface can be suitably tuned according to the input signals by simply changing the IL’s viscosity. This could pave the way for edge AI devices that can accurately learn the various signals produced in the living environment in real time.
Computing has never been more flexible!
Reference: “Reservoir computing with dielectric relaxation at an electrode–ionic liquid interface” by Sang-Gyu Koh, Hisashi Shima, Yasuhisa Naitoh, Hiroyuki Akinaga and Kentaro Kinoshita, 28 April 2022, Scientific Reports.
DOI: 10.1038/s41598-022-10152-9
Kinoshita Kentaro is a Professor at the Department of Applied Physics at Tokyo University of Science, Japan. His area of interest is device physics, with a focus on memory devices, AI devices, and functional materials. He has published 105 papers with over 1600 citations to his credit and holds a patent to his name.
This study was partly supported by JSPS KAKENHI Grant Number JP20J12046.
Tokyo University of Science (TUS) is a well-known and respected university, and the largest science-specialized private research university in Japan, with four campuses in central Tokyo and its suburbs and in Hokkaido. Established in 1881, the university has continually contributed to Japan’s development in science through inculcating the love for science in researchers, technicians, and educators.
