Lightmatter, a startup founded by physicists at MIT’s Quantum Photonics Laboratory, won a $100,000 launch grant based on a novel idea for the future of quantum computing – using light to perform calculations.
The grant, which Lightmatter won on May 17th, 2017, is part of an annual tradition held by MIT since 1990. The program is called MIT $100K, and involves separate grand prizes for teams developing innovative technology, organized into three categories – Pitch, Accelerate, and Launch.
Of these three, the Launch prize is the one awarded to teams closest to producing working prototypes of their ideas. MIT then connects the team to venture capitalists, serial entrepreneurs, corporate executives, attorneys, and startup services, giving them the foundation they need to produce new, groundbreaking products for the world at large.
Lightmatter’s Brain-Inspired Computational Model
Lightmatter’s team of quantum physicists realized that current digital computing models suffer from an important setback – they process operations one at a time. Since modern computers use electricity to identify individual bits of information, they must process data sequentially – working quickly to process large amounts of information passing through their circuits.
Light, however, operates in a fundamentally different way than electricity does. Because they exhibit wave-particle duality, photons offer an intriguing possibility – computers could use them to perform simultaneous calculations, removing the rubbernecking obstacle blocking further progress in computing speed. That extra speed is needed for high-volume data operations such as those used in artificial intelligence and machine learning.
Dr. Dirk Englund, assistant professor at MIT and one of the company’s founding members, describes the idea to us photons for computational purposes as being analogous to the operation of the human brain.
“Every neuron in your brain is connected to in the order of 7,000 other neurons. In an electrical circuit, you have lots of wires crossing over one another. Light flies through to make these highly connected networks, creating a huge advantage” — Dr. Dirk Englund, founder, Lightmatter
The potential of modeling an optical computational circuit after the human brain has remarkable implications for the development of artificial intelligence. Some experts believe that once quantum computers replace silicon for generating decision-making modeling processes, a technological artificial intelligence singularity may be reached.
What is the Technological Singularity?
Technological singularity refers to the point when computational power reaches a point beyond which meaningful prediction can no longer occur. Until now, computational advances have followed a roughly predictable trend made most famous by Moore’s Law. Transistors get smaller, cheaper, and less expensive over time, and have done so in a very regular way for the past half-century.
The singularity occurs at the point where Moore’s Law is no longer relevant – when artificially intelligent machines are capable of improving themselves and creating new machines even more intelligent than they are. Experts’ attempts at describing what a post-singularity society will look are limited – there is simply no way to predict how a hyper-intelligent artificial intelligence may choose to use its unlimited intellectual powers.
Experts such as Vernor Vinge predict that computing capable of mirroring the operational structure and efficiency of a human brain can produce the conditions necessary for the singularity event. So far, silicon and transistor-based computers have been unable to demonstrate this level of intelligence – or anything even close to it – but quantum computing might.
How Do Photons Perform Calculations?
Quantum computers rely on a unique physical phenomenon to function. This phenomenon is called superposition – it happens when a quantum particle (such as a photon) inhabits to contradictory states at the same time.
This sounds impossible, but scientists have observed this behavior in a variety of quantum physics experiments. Quantum computers use this phenomenon to create quantum bits capable of processing multiple solutions to computing problems simultaneously.
Until now, most quantum computing experiments have relied on ions trapped in oscillating magnetic fields, superconducting circuits cooled down to nearly 273 degrees Celsius, or defects in the crystal structures of diamonds. Photons offer superior calculating power compared to all of these materials, but are difficult to control.
MIT has a prototype of a device capable of confining and controlling photons exhibiting quantum superpositions. The device works by trapping two photons and entangling them using this strange physical property – if a single photon enters the device, it will simply pass through it with no effect.
While the device is currently too unstable to perform quantum computing work consistently, it represents the beginning of physicists’ upcoming work understanding the nature of quantum computing using photons and provides a starting point for light-based work that does not rely on magnetic fields, low temperatures, or expensive diamonds.
What Can Quantum Computers Do?
Apart from conjecture about brain-modeled quantum computers inheriting artificial intelligence, this technology comes with important implications. Classical computers, for instance, have great difficulty finding the prime factors of large numbers.
Most data encryption schemes rely on prime factorization problems because they are so difficult to solve – finding the prime factors of a 232-digit number requires running hundreds of classical computers in parallel for two years. A properly calibrated quantum computer, on the other hand, could theoretically solve such a problem in a fraction of that time.
Quantum computing also offers groundbreaking implications for medicine, allowing for the simulation of hypercomplex drug interactions. Understanding, simulating, and eventually controlling complex physical systems – such as the human body, the weather, or photosynthesis – will likely only become possible once Lightmatter achieves its goal of building a scalable, economically viable quantum computer.
Image Source: Adobe Stock