Nippon Telegraph and Telephone Corporation NTT's research division has conducted a collaborative project with the University of Tokyo's International Research Center for Neurointelligence or IRCN to examine and produce innovations based on coherent Ising machine technology.


Coherent Ising Machine: Neuromorphical and AI Combined

art-blue-pattern-texture-8782083
(Photo: Sergey Katyshkin from Pexels)

The two organizations will push to develop coherent Ising machine or CIM processors that use algorithms that could provide solutions to the most difficult sets of problems. CIM will serve as an information processor, similar to the function of most AI devices, and will use photonic oscillator networks. University of Tokyo's professor and IRCN Deputy Director Kazuyuki Aihara is the leader of the collaborative research with regard to artificial intelligence studies and algorithm-based models.

Another part of the team for the network computing research is Satoshi Kako, NTT's Physics & Informatics Laboratory's expert, who is responsible for observing the effectiveness and application of the coherent Ising machine-based technology.

The main interest of the study with PHI Lab and NTT is to focus on the issues experienced in utilizing the Ising model. Most of the problems that are encountered in the CIM are mathematical processing of magnetic systems that includes the controlling of fundamental particles' angular momentum and spins. Through the research, both institutes will have a chance to develop a combination of neuromorphic solutions and machine learning AI systems.

ALSO READ: Is Merging of AI and Human Life Possible with Neuralink Brain Chip?


Next-Gen Neuro-Computing System

CIM's main function is to simply give answers to problems through processing an optimal combination of different potential variables. These variables are under various sets of conditions and are one of the key components of CIM's combinatorial optimization problems. The collaborative effort of IRCN and NTT aims to discover computing principles as well as algorithms that could potentially be applied and made compatible with the CIM's modern version, specifically the digital platform. CIM should be able to simulate 16,000 spins and various couplings through the field-programmable gate array or FPGA.

PHI Lab director Yoshihisa Yamamoto, IRCN project professor Timothee Leleu, and Stanford University's experts Surya Ganguli and Hideo Mabuchi are the authors of the coherent Ising machine study published in the journal Applied Physics Letters, titled "Coherent Ising Machines-Quantum Optics and Neural Network Perspectives."

Yamamoto said that the CIM development belongs to the team's goal of innovating new scientific advancements. They were also looking forward to the application of the CIM simulator and digital algorithms to be used in contributive studies with regard to finding neuro-computing solutions and other applications that are suitable for the said project.

According to a report by AiThority, Aihara said that the complex mathematical and modeling system complements the theoretical CIM platform that could possibly be the next generation of networking, data-driven structural processing, and non-linear data analysis. Aihara's proven expertise with neuromorphic technology studies had been exhibited before from previous researches, including those in the field of biological disease prediction and artificial intelligence.

CIM technology is a significant medium to redesign modern-day artificial quantum computers, as well as the initial variants of AI-powered digital devices. To relay more comprehensive research and achieve the best out of the coherent Ising machines, the PHI Lab of the NTT research division has exchanged their interest with multiple prestigious universities, including CalTech and MIT, and a joint partnership with NASA Ames Research Center.

RELATED ARTICLE: Elon Musk's Brain Chip Could Create Universal Language in Five to Ten Years

Check out more news and information on Artificial Intelligence in Science Times.