Mediaboss Marketing

Weve painted ourselves into another corner with artificial intelligence. Were finally starting to breakthrough the usefulness barrier but were butting up against the limits of our our ability to responsibly meet our machines massive energy requirements.

At the current rate of growth, it appears well have to turn Earth into Coruscant if we want to keep spending unfathomable amounts of energy training systems such as GPT-3 .

The problem: Simply put, AI takes too much time and energy to train. A layperson might imagine a bunch of code on a laptop screen when they think about AI development, but the truth is that many of the systems we use today were trained on massive GPU networks, supercomputers, or both. Were talking incredible amounts of power. And, worse, it takes a long time to train AI.

The reason AI is so good at the things its good at, such as image recognition or natural language processing, is because it basically just does the same thing over and over again, making tiny changes each time, until it gets things right. But were not talking about running a few simulations. It can take hundreds or even thousands of hours to train up a robust AI system.

One expert estimated that GPT-3, a natural language processing system created by OpenAI, would cost about $4.6 million to train. But that assumes one-shot training. And very, very few powerful AI systems are trained in one fell swoop. Realistically, the total expenses involved in getting GPT-3 to spit out impressively coherent gibberish are probably in the hundreds-of-millions.

GPT-3 is among the high-end abusers, but there are countless AI systems out there sucking up hugely disproportionate amounts of energy when compared to standard computation models.

The problem? If AI is the future, under the current power-sucking paradigm, the future wont be green. And that may mean we simply wont have a future.

The solution: Quantum computing.

An international team of researchers, including scientists from the University of Vienna, MIT, Austria, and New York, recentlypublishedresearch demonstrating quantum speed-up in a hybrid artificial intelligence system.

In other words: they managed to exploit quantum mechanics in order to allow AI to find more than one solution at the same time. This, of course, speeds up the training process.

Per the teams paper:

The crucial question for practical applications is how fast agents learn. Although various studies have made use of quantum mechanics to speed up the agents decision-making process, a reduction in learning time has not yet been demonstrated.

Here we present a reinforcement learning experiment in which the learning process of an agent is sped up by using a quantum communication channel with the environment. We further show that combining this scenario with classical communication enables the evaluation of this improvement and allows optimal control of the learning progress.

How?

This is the cool part. They ran 10,000 models through 165 experiments to determine how they functioned using classical AI and how they functioned when augmented with special quantum chips.

And by special, that is to say, you know how classical CPUs process via manipulation of electricity? The quantum chips the team used were nanophotonic, meaning they use light instead of electricity.

The gist of the operation is that in circumstance where classical AI bogs down solving very difficult problems (think: supercomputer problems), they found thehybrid-quantum system outperformed standard models.

Interestingly, when presented with less difficult challenges, the researchers didnt not observe anyperformance boost. Seems like you need to get it into fifth gear before you kick in the quantum turbocharger.

Theres still a lot to be done before we can roll out the old mission accomplished banner. The teams work wasnt the solution were eventually aiming for, but more of a small-scale model of how it could work once we figure out how to apply their techniques to larger, real problems.

You can read the whole paper here on Nature.

H/t: Shelly Fan, Singularity Hub

Published March 17, 2021 19:41 UTC

See the article here:
The key to making AI green is quantum computing - The Next Web

Author: David Orrell, Author and Economist

March 16, 2021

In 2019, Google announced that they had achieved quantum supremacy by showing they could run a particular task much faster on their quantum device than on any classical computer. Research teams around the world are competing to find the first real-world applications and finance is at the very top of this list.

However, quantum computing may do more than change the way that quantitative analysts run their algorithms. It may also profoundly alter our perception of the financial system, and the economy in general. The reason for this is that classical and quantum computers handle probability in a different way.

The quantum coinIn classical probability, a statement can be either true or false, but not both at the same time. In mathematics-speak, the rule for determining the size of some quantity is called the norm. In classical probability, the norm, denoted the 1-norm, is just the magnitude. If the probability is 0.5, then that is the size.

The next-simplest norm, known as the 2-norm, works for a pair of numbers, and is the square root of the sum of squares. The 2-norm therefore corresponds to the distance between two points on a 2-dimensional plane, instead of a 1-dimensional line, hence the name. Since mathematicians love to extend a theory, a natural question to ask is what rules for probability would look like if they were based on this 2-norm.

It is only in the final step, when we take the magnitude into account, that negative probabilities are forced to become positive

For one thing, we could denote the state of something like a coin toss by a 2-D diagonal ray of length 1. The probability of heads is given by the square of the horizontal extent, while the probability of tails is given by the square of the vertical extent. By the Pythagorean theorem, the sum of these two numbers equals 1, as expected for a probability. If the coin is perfectly balanced, then the line should be at 45 degrees, so the chances of getting a heads or tails are identical. When we toss the coin and observe the outcome, the ambiguous state collapses to either heads or tails.

Because the norm of a quantum probability depends on the square, one could also imagine cases where the probabilities were negative. In classical probability, negative probabilities dont make sense: if a forecaster announced a negative 30 percent chance of rain tomorrow, we would think they were crazy. However, in a 2-norm, there is nothing to prevent negative probabilities occurring. It is only in the final step, when we take the magnitude into account, that negative probabilities are forced to become positive. If were going to allow negative numbers, then for mathematical consistency we should also permit complex numbers, which involve the square root of negative one. Now its possible well end up with a complex number for a probability; however the 2-norm of a complex number is a positive number (or zero). To summarise, classical probability is the simplest kind of probability, which is based on the 1-norm and involves positive numbers. The next-simplest kind of probability uses the 2-norm, and includes complex numbers. This kind of probability is called quantum probability.

Quantum logicIn a classical computer, a bit can take the value of 0 or 1. In a quantum computer, the state is represented by a qubit, which in mathematical terms describes a ray of length 1. Only when the qubit is measured does it give a 0 or 1. But prior to measurement, a quantum computer can work in the superposed state, which is what makes them so powerful.

So what does this have to do with finance? Well, it turns out that quantum algorithms behave in a very different way from their classical counterparts. For example, many of the algorithms used by quantitative analysts are based on the concept of a random walk. This assumes that the price of an asset such as a stock varies in a random way, taking a random step up or down at each time step. It turns out that the magnitude of the expected change increases with the square-root of time.

Quantum computing has its own version of the random walk, which is known as the quantum walk. One difference is the expected magnitude of change, which grows much faster (linearly with time). This feature matches the way that most people think about financial markets. After all, if we think a stock will go up by eight percent in a year then we will probably extend that into the future as well, so the next year it will grow by another eight percent. We dont think in square-roots.

This is just one way in which quantum models seem a better fit to human thought processes than classical ones. The field of quantum cognition shows that many of what behavioural economists call paradoxes of human decision-making actually make perfect sense when we switch to quantum probability. Once quantum computers become established in finance, expect quantum algorithms to get more attention, not for their ability to improve processing times, but because they are a better match for human behaviour.

Link:
Quantum computing is finally having something of a moment - World Finance

Munich and Paris, March 18, 2021 Atos today announced that it has delivered its Atos Quantum Learning Machine (Atos QLM), the world's highest-performing commercially available quantum simulator, to the Leibniz Supercomputing Centre (LRZ), of the Bavarian Academy of Sciences and Humanities. The Atos QLM is installed in the recently opened LRZ Quantum Integration Centre (QIC), Bavarias preeminent computing facility. The center was designed to bring practical quantum applications to the scientific community by advancing the convergence of quantum computing and supercomputing.

The LRZ is among the first computing centers worldwide to focus on the integration of quantum computing in an HPC environment with its Quantum Integration Centre. The hybrid quantum-HPC approach shows significant promises in effectively using todays classical computers to harness the power of near-term quantum applications. Leveraging both the Atos QLM and its collaboration with key players like Atos, the Finnish-German startup IQM and other partners, LRZ will be able to make quantum technologies available to more users. By taking advantage of existing HPC infrastructures, this initiative will allow them to explore and capture the opportunities made possible by quantum computing within a couple of years.

At the LRZ, we are a partner for digitalization in science. We are expanding our portfolio by integrating services for quantum computing. This way we enable world-class researchers to find new approaches to solving grand-challenge scientific problems. However, we are only at the beginning with this technology. At the LRZ Quantum Integration Centre, scientists will be able to learn how to use it and prepare themselves for the future of quantum computing. The collaboration with Atos and the use of the Atos Quantum Learning Machine are an essential building block in our Quantum Computing strategy, explained Prof. Dieter Kranzlmller, Chairman of the Leibniz Supercomputing Centre.

LRZ and Atos share a very pragmatic approach to quantum computing that focuses on quantum-accelerated HPC, with the aim of delivering early strategic benefits to users before we fully enter the post-quantum era. The Atos QLM is a direct extension of this approach and we are honored to be one of the first hardware partners of the LRZ Quantum Integration Centre. It is a fantastic project and marks the significant contribution made by LRZ to the quantum computing community, said Elie Girard, Atos CEO.

The LRZ Quantum Integration Centre supports the Munich Quantum Valley, a central element of the Bavarian quantum initiative to drive quantum computing forward at a national and international level. The partnership between Atos and LRZ is a testament to the ambition of the Bavarian authorities to become an internationally competitive quantum location by incorporating international, leading-edge knowledge, skills and technologies. Subject to the approval of the state parliament, the Free State of Bavaria committed to providing a total of 300 million euros.

***

About AtosAtos is a global leader in digital transformation with 105,000 employees and annual revenue of over 11 billion. European number one in cybersecurity, cloud and high performance computing, the Group provides tailored end-to-end solutions for all industries in 71 countries. A pioneer in decarbonization services and products, Atos is committed to a secure and decarbonized digital for its clients. Atos operates under the brands Atos and Atos|Syntel. Atos is a SE (Societas Europaea), listed on the CAC40 Paris stock index.

The purpose of Atos is to help design the future of the information space. Its expertise and services support the development of knowledge, education and research in a multicultural approach and contribute to the development of scientific and technological excellence. Across the world, the Group enables its customers and employees, and members of societies at large to live, work and develop sustainably, in a safe and secure information space.

Press contact:Marion Delmas | marion.delmas@atos.net | +33 6 37 63 91 99

See the article here:
Atos supports the Leibniz Supercomputing Centre in pioneering quantum-accelerated computing with the Atos QLM - GlobeNewswire

Photograph of a quantum computing chip that a Google team used in their claimed quantum computer. Photo: Nature 574, 505-510 (2019).

The Union finance ministry presented the national budget for 2021 one and a half months ago. One of the prime motivations of a nationalist government should be cyber-security, and it is high time we revisited this technological space from the context of this budget and the last one.

One of the highlights of the 2020 budget was the governments new investment in quantum computing. Finance minister Nirmala Sitharamans words then turned the heads of researchers and developers working in this area: It is proposed to provide an outlay of 8,000 crore rupees over a period of five years for the National Mission on Quantum Technologies and Applications.

Thanks to the pandemic, it is not clear how much funding the government transferred in the first year. The 2021 budget speech made no reference to quantum technologies.

Its important we discuss this topic from a technological perspective. Around four decades ago, physicist Richard Feynman pointed out the possibility of devices like quantum computers in a famous speech. In the early 1990s, Peter Shor and others proved that such computers could easily factor the product of two large prime numbers a task deemed very difficult for the classical computers we are familiar with. This problem, of prime factorisation, underlies the utility of public key crypto-systems, used to secure digital transactions, sensitive information, etc. online.

If we have a practicable quantum computer, the digital security systems currently in use around the world will break down quickly, including that of financial institutions. But commercial quantum computers are still many years away.

On this count, the economically developed nations are on average far ahead of others. Countries like the US, Canada, Australia and China have already made many advancements towards building usable quantum computers with meaningful capabilities. Against this background, the present governments decision in February 2020 to invest such a large sum in quantum technologies was an outstanding development.

The problem now lies with distributing the money and achieving the actual technological advances. So far, there is no clear evidence of this in the public domain.

A logical step in this direction would be to re-invest a large share of the allocation in indigenous development. This is also where the problems lie. One must understand that India has never been successful in fabricating advanced electronic equipment. While we have very good software engineers and theoretical computer scientists, there is no proven expertise in producing chips and circuits. We might have some limited exposure in assembling and testing but nothing beyond that.

So while Atmanirbhar Bharat is an interesting idea, it will surely take a very long time before we find ourselves able to compete with developed nations vis--vis seizing on this extremely sophisticated technology involving quantum physics. In the meantime, just as we import classical computers and networking equipment, so should we proceed by importing quantum equipment, until our indigenous capability in this field matures to a certain extent.

For example, demonstrating a four-qubit quantum system or designing a proof-of-concept quantum key distribution (QKD) circuit might be a nice textbook assignment. However, the outcome will not nearly be competitive to products already available in the international arena. IBM and Google have demonstrated the use of machines with more than 50 qubits. (These groups have participation from Indian scientists working abroad.) IBM has promised a thousand-qubit machine by 2023. ID Quantique has been producing commercial QKD equipment for more than five years.

India must procure such finished products and start testing them for security trapdoors before deploying them at home. Doing so requires us to train our engineers with state-of-the-art equipment as soon as possible.

In sum, indigenous development shouldnt be discontinued but allocating a large sum of money for indigenous development alone may not bring the desired results at this point.

By drafting a plan in the 2020 Union budget to spend Rs 8,000 crore, the government showed that it was farsighted. While the COVID-19 pandemic has made it hard to assess how much of this money has already been allocated, we can hope that there will be renewed interest in the matter as the pandemic fades.

This said, such a huge allocation going to academic institutes and research laboratories for trivial demonstrations might be imprudent. In addition, we must begin by analysing commercially available products, made by international developers, so we can secure Indias security infrastructure against quantum adversaries.

Serious science requires deep political thought, people with strong academic commitment in the government and productive short- as well as long-term planning. I hope the people in power will enable the Indian community of researchers to make this quantum leap.

Subhamoy Maitra is a senior professor at the Indian Statistical Institute, Kolkata. His research interests are cryptology and quantum computing.

Read more:
After the Govt's Big Allocation on Quantum Technologies in 2020, What Next? - The Wire Science

The Colorado Economic Development Commission normally doesnt throw its weight behind unproven startups, but it did so on Thursday, approving $2.9 million in state job growth incentive tax credits to try and land a manufacturing plant that will produce hardware for quantum computers.

Given the broad applications and catalytic benefits that this companys technology could bring, retaining this company would help position Colorado as an industry leader in next-generation and quantum computing, Michelle Hadwiger, the deputy director of the Colorado Office of Economic Development & International Trade, told commissioners.

Project Quantum, the codename for the Denver-based startup, is looking to create up to 726 new full-time jobs in the state. Most of the positions would staff a new facility making components for quantum computers, an emerging technology expected to increase computing power and speed exponentially and transform the global economy as well as society as a whole.

The jobs would carry an average annual wage of $103,329, below the wages other technology employers seeking incentives from the state have provided, but above the average annual wage of any Colorado county. Hadwiger said the company is also considering Illinois, Ohio and New York for the new plant and headquarters.

Quantum computing is going to be as important to the next 30 years of technology as the internet was to the past 30 years, said the companys CEO, who only provided his first name Corban.

He added that he loves Colorado and doesnt want to see it surpassed by states like Washington, New York and Illinois in the transformative field.

If we are smart about it, and that means doing something above and beyond, we can win this race. It will require careful coordination at the state and local levels. We need to do something more and different, he said.

The EDC also approved $2.55 million in job growth incentive tax credits and $295,000 in Location Neutral Employment Incentives for Nextworld, a growing cloud-based enterprise software company based in Greenwood Village. The funds are linked to the creation of 306 additional jobs, including 59 located in more remote parts of the state.

But in a rare case of dissent, Nextworlds CEO Kylee McVaney asked the commission to go against staff recommendations and provide a larger incentive package.

McVaney, daughter of legendary Denver tech entrepreneur Ed McVaney, said the companys lease is about to expire in Greenwood Village and most employees would prefer to continue working remotely. The company could save substantial money by not renewing its lease and relocating its headquarters to Florida, which doesnt have an income tax.

We could go sign a seven-year lease and stay in Colorado or we can try this new grand experiment and save $11 million, she said.

Hadwiger insisted that the award, which averages out to $9,500 per job created, was in line with the amount offered to other technology firms since the Colorado legislature tightened the amount the office could provide companies.

But McVaney said the historical average award per employee was closer to $18,000 and the median is $16,000 and that Colorado was not competitive with Florida given that states more favorable tax structure.

See the article here:
Colorado makes a bid for quantum computing hardware plant that would bring more than 700 jobs - The Denver Post