The neuromorphic system uses quantum effects to find optimal solutions to complex problems

It is easy to solve a Rubik’s Cube 3×3, explains Shantanu Chakrabarty, Professor of Clifford W. Murphy and Vice-Dean of Research and Higher Studies at McKelvey School of Engineering at Washington University in St. Louis. Learn and memorize the steps, then run them to achieve the solution.

Computers are already good in this type of procedural problem solving. Now Chakrabarty and its employees have developed a tool that can go beyond the procedure to discover new solutions to complex logistics optimization problems to discover drugs.

Chakrabarty and its collaborators introduced Neurosa, a neuromorphic architecture of problem solving modeled on the functioning of human neurobiology, but which exploits quantum mechanical behavior to find optimal – guaranteed – solutions and find these solutions more reliable than advanced methods.

The effort of multi-university collaboration, published in Nature communicationsOriginally from the Neuromorphic and Cognitive Engineering workshop in Telluride and was led by Chakrabarty and the first author Zihao Chen, a graduate student from the Green Department of Preston M. Green of electric engineering and systems in McKelvey Engineering.

“We are looking for ways to solve problems better than modeled computers on human learning have already done,” said Chakrabarty. “Neurosa is designed to solve the problem of” discovery “, the most difficult problem of automatic learning, where the objective is to discover new unknown solutions.”

In optimization, the receipt is a process to explore different possible solutions before finally settled on the best solution. The Rectors of Fowler-Nordheim (FN) use principles of quantum mechanical tunneling to seek this most optimal solution effectively, and they are “the secret ingredient” in Neurosa, known as Chakrabarty.

“In optimization problems, the strategy comes into play when the system has to change, like when you are looking for the highest building in campus, when do you move to another area?” Chakrabarty said. “The structure of Neurosa is neuromorphic, like our brain structure with neurons and synapses, but its research behavior is determined by the FN Benaler. This critical bridge between Neuro and Quantum is what makes Neurosa so powerful and what allows us to guarantee that we will find a solution if we give enough time.”

This guarantee becomes particularly important when the calendar to allow Neurosa to search for an optimal solution could go from days to weeks, or even more, depending on the complexity of the problem.

In the article, the Chakrabartty team, in collaboration with a research team at Spinncloud Systems, has already demonstrated that Neurosa can be implemented on the neuromorphic SPINNAKER2 computer platform, proving its practical feasibility. Then, Chakrabartty plans that the tool could be applied to the optimization of logistics in supply chains, manufacturing and transport services or to discover new drugs by exploring the retreat of proteins and optimal molecular configurations.