Thursday, October 6, 2022

IBM’s Potential Quantum AI Advantage

IBM has two moonshot efforts that are potentially game changing for both the company and the industry: Watson and related artificial intelligence (AI) efforts; and the company’s quantum computing initiative. 

They are both fascinating. It’s reported that Watson’s use as a conversational computer has been so successful in trial that the implementations are often asked out on dates. IBM’s quantum development, which is well in advance of when quantum computers will be in general use, is focused on building the quantum infrastructure so that when quantum computers show up, the industry will be ready.

But things get interesting when we talk about blending the two technologies and using quantum computers to make AIs far more capable and accurate.

Let’s talk about this coming integration between the two technologies:

The problem with deep learning

The best way to assure the accuracy and capability of an AI is through deep learning (DL). 

Deep learning allows the AI to develop at its own speed with direct access to massive amounts of data. This allows the AI to self-train at scale. The problem is that the amount of data needed and the ability of the AI to consume that data is prohibitive to a successful and timely deployment of the resulting system.

So machine learning (ML), where the AI is trained by humans, is still more common, even though the result isn’t as accurate or as reliable. 

Quantum’s advantage

Quantum computers have one big advantage, which is the ability to deal with and make determinations from massive datasets.  

If the quantum computer were added to the AI’s training process, it could synthesize the massive datasets into relatively accurate summaries that then could be fed to the deep learning AI. This would reduce both the amount of data needed to train the AI and reduce potential errors if sampling had been used instead.  

Quantum computers could also break out anomalies, passing them through to the deep learning computer, so it would be trained on the exceptions.  

The result should be more common use of deep learning coupled with a quantum computer front end to synthesize the data and ensure a better, faster outcome.

IBM’s potential advantage

IBM is one of the companies working on both technologies, and thus far, it’s more likely than any Western company to craft a blended outcome. 

This blended outcome of AI and quantum computing would be invaluable to most broad uses of AI and help assure the West has a competing offering to those expected out of China. 

If IBM can execute on these initiatives, and the company has a strong history of executing on projects at this level, by the 2030 timeframe, it should result in what could be a significant competitive advantage to IBM’s AI efforts, as it moves beyond conversational computing and into quantum AI.  

Quantum AI computers will be in a class by themselves. They’ll be able to address some of the largest and most complex problems and decisions faced by the world and provide relatively unbiased and accurate and complete answers. 

Quantum AI

Both AI and quantum computers are game-changing technologies. When we blend the two concepts together, we’ll be able to increasingly interact with computers as we interact with each other — and get insights into the most complex and data-intensive problems in the universe.

When this technology matures, likely in the 2030s, those companies already up to speed on both technologies will undoubtedly be the biggest beneficiaries. There’s a decent chance that this timeline will get shorter rather than longer, as more and more companies pivot toward this opportunity. 

Similar articles

Latest Articles