Leverage Quantum Computing and AI for Self-reliance in Semiconductor: SAPI

Leveraging Quantum Computing and AI will help us realising our dream of self-reliance in Semiconductor, said SAPI
Leveraging Quantum Computing and AI will help us realising our dream of self-reliance in Semiconductor, said SAPI
Published on

New Delhi: Delhi-based Think Factory 'Security, Peace and policy Initiatives' (SAPI) organized the second webinar of the 'Demystifying AI' series on Sunday, July 24. The topic of the webinar was 'Quantum Computing and AI for self-reliance in Semiconductors for India'. The webinar was so successful that the distinguished scientist of DRDO and founder-CEO of Brahmos Aerospace Pvt Ltd Dr Sivathanu Pillai himself lauded the unique idea of simulation to design better semiconductors using Quantum Computers, suggested by Deep Prasad, CEO of Vancouver-based startup 'Quantum Generative Materials' (GenMat).

Deep Prasad was the keynote speaker in the webinar. GenMet works on applications of Quantum Computing and AI and has recently raised USD 50 million for its key projects.

Apart from Dr Sivathanu Pillai, the webinar was attended by veteran scientists & researchers from India and abroad, founders of startups, and members of the industry and academia. The founders of SAPI Pradeep Gupta and Sarang Nerkar introduced the subject and highlighted its relevance and importance to catch up in the race of technology.

"Leveraging Quantum Computing and AI to speed-up simulations"

While elaborating on Simulation Model, Deep stated that we must build a digital twin of semiconductors (it is a virtual replica), but doing so requires being able to model quantum phenomena (electron tunnelling, excitons, etc). This requires building a physics-based model through a number of computational materials design techniques namely Density Functional Theory, Post Density Functional Theory, Molecular Dynamics, etc. All of these simulations take a long time to run (several days on average). He stated that we can speed up these simulations by training an AI to calculate the solution to the underlying equations that are being solved by the simulations.

Deep further highlighted that 'GenMat' has a 'Materials Discovery Platform' which takes 1 day and 7 hours on average to complete a simulation. Each family of simulations would cost hundreds of thousands to millions of dollars to elucidate experimentally. To virtually map, any physical entity will require a huge number of simulations. Here 'GenMat' is trying to reduce this simulation time even further using AI, down to hours in the near term and eventually seconds.

How this will be achieved?

Deep suggested using Quantum Computers (QC) which have very high computational power than the regular computers which we use in our daily lives. The QC is based on principles of quantum mechanics to represent, store and compute quantum information thru 'Qubits' ( which have infinite possibilities) while regular computers are based on principles of classical mechanics to represent, store and compute classical information done in binary bits ( either 0 or 1). Quantum Mechanics better explains the real world around us which is probabilistic and unpredictable while classical physics which is deterministic fails to forecast all noticeable phenomena occurring in nature. The world we observe at the macroscopic level feeds us classical information, which has physical rules and intuition we understand. The world at the microscopic level (sub-atomic levels) consists of quantum information, which has physical rules and intuition that are alien.

Therefore Deep pitched using quantum computers to compute quantum information. Quantum information includes the quantum physical properties of semiconductors, which require supercomputers to simulate very imperfectly. By capturing the true quantum physics behind semiconductors with quantum computers, we can properly simulate the behaviour of semiconductors and determine how to optimize them without experimentally building as many or as often. We can then train an AI to make key optimization design decisions for the semiconductor based on the quantum computing results.

What can India do about it?

Deep had the following recommendations as a road map for what India can do:-
• Invest in quantum computing, artificial intelligence, advanced manufacturing and materials engineering technologies
• Build relevant global partnerships with private and government institutions that are aligned with democratic principles
• Take a simulation-first approach to build the next generation of semiconductor research and development, as well as fabrication facilities
• Start small, end big (take existing fabrication tech that is cheap, optimize it using AI and Quantum)
• Have a ten-year roadmap, with AI to deliver the primary value for the first 5 years, and Quantum to deliver the primary value thereafter

Dr Sivathanu Pillai lauded this unique idea of simulation to design better semiconductors using Quantum Computers. He was highly appreciative of the same and suggested that SAPI should connect with the Electronics Industry in Bangalore and organize sessions to further carry forward this novel idea. He also supported the initiative of the Global Partnership.

Further, the ensuing discussion of using QC in other fields like the use of Hyperspectral sensors in the Satellites in agriculture will go long way. He highlighted the newer application of Precision Farming to better the Farming Practices with available resources.

(Defence Watch– India's Defence News centre that focuses on Defence Manufacturing, Defence Technology, Strategy and Military affairs is on Twitter. Follow us here and stay updated.)

logo
Defence Watch
www.defencewatch.in