open share links close share links

About 92% of large companies are achieving returns on their investments in artificial intelligence, and the same percentage are increasing their AI investments. But what does it take for startups and early-stage companies to get to this point?

Thats a critical question, according to Sukwoong Choi, a postdoctoral scholar at MIT Sloan. AI utilization is tied to startups products and services. Its more directly relevant, he said.

In a newpaper, Choi and his co-authors find that firms need to be ready to make a significant investment in AI to see any gains, because limited AI adoption doesnt contribute to revenue growth. Only when firms increase their intensity of AI adoption to at least 25% meaning that they are using a quarter of the AI tools currently available to them do growth rates pick up and investments in AI start to pay off.

The paper was co-authored by Yong Suk Lee, Taekyun Kim, and Wonjoon Kim.

Here are three things companies should know about investing in AI.

The researchers surveyed 160 startups and small businesses in South Korea about their use of AI technologies such as natural language processing, computer vision, and machine learning. Of the firms included, 53% were in technology-related fields (namely software, pharma, and mobile computing), and 54% had adopted AI to some degree.

The survey was administered to companies created before 2015, as these firms were founded before AI adoption generally took off in South Korea. (A footnote in the paper points to an explosion of interest in AI in the country after Go master Lee Sedol lost four of five matches to Google DeepMinds AlphaGo program in March 2016.)

Among the firms surveyed, the correlation between AI adoption and revenue growth followed a J-curve: slow and steady at first, then substantial. The turning point was an intensity of AI adoption of 25%. For firms with AI intensity below 25%, annual revenue growth was essentially zero; for firms above the 25% threshold, growth approached 24%.

Theres a disruptive power for AI. With lower utilization, its harder to make a profit, Choi said. When youre in those early stages of AI adoption, you may need some time to obtain the payoff to using AI.

Several factors can influence a firms embrace of AI, the researchers found. For example, firms that are smaller and/or were founded by CEOs with prior entrepreneurial experience are more likely to adopt AI intensively. Larger firms or spinoffs from other companies are less likely to adopt AI at that level, though lab-based spinoffs are an exception.

One of the most influential factors, though, is adoption of complementary technology namely, big data capabilitiesand cloud computing. The former contributes to better AI outcomes through more mature data collection and management, while the latter provides the computational power necessary to run complex analyses. Both help firms drive growth from their investments in AI.

This finding came as little surprise to Choi and his co-authors. For decades, investing in one type of technology has driven the adoption of other technologies. Examples abound: Better operating systems led to better software, faster modems made computer networks possible, and IT infrastructure supported the growth of online selling.

Complementary technology makes it easy to adopt new technology such as AI, Choi said. To adopt and utilize AI effectively, and to get the payoff at earlier stages in your investment, you need the technology and the skills that go with it.

The pivotal role of complementary technology points to one key takeaway from the paper, Choi said. To support AI adoption, its not enough to have access to the technology you also need the infrastructure that supports it. When you make that easily available, you can accelerate AI adoption, Choi said.

The second consideration is how closely AI is tied to a companys core product or service, he said, and how that impacts the companys research and development strategy.

Internally focused R&D helps a company build absorptive capacity in this case, AI know-how that positions it to more intensively adopt and use AI technology. This is helpful for firms that need to protect their proprietary algorithms as intellectual property, or for firms working with sensitive data sets theyd rather not allow a third party to process.

On the other hand, if AI is a complement to the work that a firm is doing but isnt the core focus of that work, firms can turn to external resources, Choi said. Large language models, such as OpenAIs ChatGPT, are a good example of this: Theyre readily available, widely used, and constantly being refined.

Its important to ask, Is there a point solution for the AI work Im trying to do? Choi said. If your area of work is more systematic, then you dont necessarily need an internally focused R&D strategy. You can license something thats already available.

Read next: how to prepare for the AI productivity boom

See original here:
Artificial intelligence pays off when businesses go all in - MIT Sloan News

The AlphaGo Zero Cheat Sheet (high-res link below)

Download the AlphaGo Zero cheat sheet

Recently Google DeepMind announced AlphaGo Zero an extraordinary achievement that has shown how it is possible to train an agent to a superhuman level in the highly complex and challenging domain of Go, tabula rasa that is, from a blank slate, with no human expert play used as training data.

It thrashed the previous reincarnation 1000, using only 4TPUs instead of 48TPUs and a single neural network instead of two.

The paper that the cheat sheet is based on was published in Nature and is available here. I highly recommend you read it, as it explains in detail how deep learning and Monte Carlo Tree Search are combined to produce a powerful reinforcement learning algorithm.

Hopefully you find the AlphaGo Zero cheat sheet useful let me know if you find any typos or have questions about anything in the document.

If you would like to learn more about how our company, Applied Data Science develops innovative data science solutions for businesses, feel free to get in touch through our website or directly through LinkedIn.

and if you like this, feel free to leave a few hearty claps 🙂

Applied Data Science is a London based consultancy that implements end-to-end data science solutions for businesses, delivering measurable value. If youre looking to do more with your data, lets talk.

Visit link:
AlphaGo Zero Explained In One Diagram | by David Foster - Medium

What is the most important product that comes out of the semiconductor industry?

Here is a hint: It is inherent to the market, but enhanced by a positively reinforcing feedback loop of history. Here is another hint: You cant hold it in your hand, like an A0 stepping of a device, and you cant point at it like a foundry with the most advanced manufacturing processes created from $15 billion to $20 billion worth of concrete, steel, and wafer etching equipment and a whole lotta people in bunny suits.

No, the most important thing that the semiconductor industry delivers and has consistently delivered for over five decades is optimism. And unlike a lot of chips these days, there is no shortage of it despite the serious challenges that the industry is facing.

By optimism we do not mean the kind of future poisoning that company founders and chief executives sometimes succumb to when they spend too much time in the future that is not yet here without seeing the consequences of the technologies they are in the process of inventing. And we certainly do not mean the zeal that others exhibit when they think that information technology can solve all of our problems. It cant, and it often makes some things worse as it is making other things better, as all technologies have done since humanity first picked up a stick. It is the arm that swings the stick both ways to plant a seed or to crush a skull. So it is with the Internet, social media, artificial intelligence, and so on.

The optimism that we are speaking of in the semiconductor industry is usually stripped bare of such consequences, with the benefits all emphasized and the drawbacks mostly ignored except possibly when considering the aspects of climate change and how compute, storage, and networking are an increasingly large part of our lives, and something that represents an ever-enlargening portion of business and personal budgets and consequently an embiggening part of the energy consumption on the planet. Semiconductor makers turn this drawback more computers requiring more power and cooling into a cause for driving innovation as hard as it can be done.

The irony is that we will need some of the most power-hungry systems the world has ever seen to simulate the conditions that will prove how climate change will affect us collectively and here is the important bit individually. How will you feel when you can drill down into a simulation, for a modest fee of course, and see a digital twin of your home being destroyed by a predicted hurricane two years from now? Or an earthquake, or a fire, or a tsunami? What is true of the Earth simulation will be as true for your body simulation and your consequent healthcare.

If the metaverse means anything, it means using HPC and AI to make general concepts extremely personal. We dont know that the world was hell bent to adopt the 24 hour news cycle and extreme entertainment optionality of cable television, or the Web, or social networks, but what we do know is that most of us ended up on these platforms anyway. And what seems clear is that immersive, simulated experiences are going to be normalized, are going to be a tool in all aspects of our lives, and that the race is on to develop the technologies that will get us there.

It would be hard to find someone more genuine and more optimistic about the future of the semiconductor industry than Aart de Geus, co-founder, chief executive officer, and chairman of electronic design automation tool maker Synopsys, who gave the opening keynote at the ISSCC 2022 chip conference, which was hosted online this week. We read the paper that de Geus presented and watched the keynote as well, and will do our best to summarize the tour de force in semiconductor history and prognostication as we enter in what de Geus called the SysMoore Era the confluence of Moores Law ambitions in transistor design and now packaging coupled to systemic complexity that together will bring about a 1,000X increase in compute across devices and systems of all kinds and lead to a smart everything world.

Here is de Geus showing the well familiar exponential plot of the transistor density of CPUs, starting with the Intel 4004 in 1971 and running all the way out five decades later to the Intel Ponte Vecchio GPU complex, with 47 chiplets lashing together 100 billion transistors, and the Cerebras WSE 2 wafer-scale processor, with 2.5 trillion transistors.

Thats the very familiar part of the SysMoore Era, of course. The Sys part needs a little explaining, but it is something that we have all been wrestling with in our next platforms. Moores Law improvements of 2X transistor density are taking bigger leaps to stay on track and are not yielding a 2X lowering in the cost of the transistors. This latter bit is what actually drives the semiconductor industry (aside from optimism), and we are now entering a time when the cost of transistors could rise a little with each generation, which is why we are resorting to chiplets and advanced packaging to glue them together side-by-side with 2.5D interposers or stacking them up in 3D fashion with vias or in many cases, a mix of the two approaches. Chiplets are smaller and have higher yield, but there is complexity and cost in the 2.5D and 3D packaging. The consensus, excepting Cerebras, is that this chiplet approach will yield the best tech-onomic results, to use a term from de Geus.

With SysMoore, we are moving from system on chip designs to system of chips designs, illustrated below, to bend up the semiconductor innovation curve that has been dominated by Moores Law for so long (with some help from Dennard scaling until 2000 or so, of course). Like this:

The one thing that is not on the charts that de Geus showed in the keynote, and that we want to inject as an idea, is that compute engines and other kinds of ASICsare definitely going to get more expensive even if the cost of packing up chiplets or building wafer-scale systems does not consume all of the benefits from higher yield that comes from using gangs of smaller chips or adding lots of redundancy into a circuit and never cutting it up.

By necessity, as the industry co-designs hardware and software together to wring the most performance per dollar per watt out of a system, we will move away from the volume economics of mass manufacturing. Up until now, a compute engine or network ASIC might have hundreds of thousands to millions of units, driving up yields over time and driving down manufacturing cost per unit. But in this SysMoore Era, volumes for any given semiconductor complex will go down because they are not general purpose, like the X86 processor in servers and PCs or the Arm system on chip was for smartphones and tablet have both been for the past decade and a half. If volumes per type of device go down by an order of magnitude, and the industry needs to make more types devices, this will put upward pressure on unit costs, too.

So what is the answer to these perplexing dilemmas that the semiconductor industry is facing? Artificial intelligence augmenting human expertise in designing these future system of chips complexes, of course. And it is interesting that the pattern that evolved to create machine learning for data analytics is being repeated in chip design.

EDA is relatively simple conceptually, explains de Geus. If you can capture data, you may be able to model it. If you can model it, maybe you can simulate. If you can simulate, maybe you can analyze. If you can analyze, maybe you can optimize. And if you can optimize, maybe you can automate. Actually, lets not forget the best automation is IP reuse it is the fastest, most efficient kind. Now its interesting to observe this because if you look at the bottom layers, what we have been doing in our field really for 50 years, is we have built digital twins of the thing that we are still building. And if we now say were going to deliver to our customers and the world that 1,000X more capability in chips, the notion of Metaverse some call it Omniverse, Neoverse, whatever you want to call it is becoming extremely powerful because it is a digital view of the world as a simulation of it.

The complexity that comprises a modern chip complex, full of chiplets and packaging, is mind-numbing and the pressure to create the most efficient implementation, across its many possible variations, is what is driving the next level of AI-assisted automation. We are moving from computer-aided design, where a workstation helped a chip designer, to electronic design automation, where synthesis of logic and the placing and routing of that logic and its memories and interconnects, is done by tools such as those supplied by Synopsys, to what we would call AIDA, short for Artificial Intelligence Design Automation, and making us think of Ada Lovelace, of course, the programmer on the Difference Engine from Charles Babbage.

This chart captures the scale of complexity in an interesting way, since the bottom two have been automated by computers IBMs Deep Blue using brute force algorithms to play chess and Googles AlphaGo using AI reinforcement learning to play Go.

Google has been using lessons learned from AlphaGo to do placement and routing of logic blocks on chips, as we reported two years ago from ISSCC 2020, and Synposys is embedding AI in all parts of its tool stack in something it is calling Design Space Optimization, or DSO. A chess match has a large number of possible moves, and Go has orders of magnitude more, but both are win-loss algorithms. Not so for route and placement of logic blocks or the possible ways to glue compute complexes together from myriad parts. These are not zero sum algorithms, but merely better or worse options, like going to the eye doctor and sitting behind that annoying machine with all the blasted lenses.

The possible combinations of logic elements and interconnects is a very large data space, and will itself require an immense amount of computation to add AI to the design stack. The amount has been increasing on a log scale since the first CAD tools became widely used:

But the good news is that the productivity gains from chip design tools have been growing at a log scale, too. Which means what you can do with one person and one workstation designing a chip is amazing here in the 2020s. And will very likely be downright amazing in the 2030s, if the vision of de Geus and his competitors comes to pass.

In the chart above, the Fusion block is significant, says de Geus, and it is implemented in something called the Fusion Compiler in the Synopsys toolchain, and this is the foundation for the next step, which is DSO. Fusion plugs all of these different tools together to share data as designers optimize a chip for power, performance, and area or PPA, in the lingo. These different tools work together, but they also fight, and they can be made to provide more optimal results than using the tools in a serial manner, as this shows:

The data shown above is an average of more than 1,000 chip designs, spanning from 40 nanometers down to 3 nanometers. With DSO, machine learning is embedded in all of the individual elements of the Fusion Compiler, and output from simulations is used to drive machine learning training that in turn is used to drive designs. The way we conceive of this and de Geus did not say this is that the more the Synopsys tools design chips and examine options in the design space, the faster it will learn what works and what does not and the better it will be at showing human chip designers how to push their designs.

Lets show some examples of how the early stages of DSO works with the Synopsys tools, beginning with a real microcontroller from a real customer:

De Geus highlighted the important parts of the design, with a baseline of the prior design and the target of the new design. A team of people were set loose on the problem using the Synopsys tools, and you can see that they beat the customer target on both power and timing by a little bit. Call it a day. But then Synopsys fired up the Fusion Compiler and its DSO AI extensions. Just using the DSO extensions to Fusion pushed the power draw down a lot and to the left a little, and then once AI trained algorithms were kicked on, the power was pushed down even further. You can see the banana curve for the DSO and DSO AI simulations, which allows designers to trade off power and timing on the chip along those curves.

Here is another design run that was done for an actual CPU as it was being designed a year ago:

A team of experts took months to balance out the power leakage versus the timing in the CPU design. The DSO extensions to the Fusion Compiler pushed it way over to the left and down a little, and when the AI trained models of the tool were switched on, a new set of power leakage and timing options were shown to be possible. A single engineer did the DSO design compared to a team using the Synopsys tools, and that single engineer was able to get a design that burned from 9 percent to 13 percent less power and had 30 percent less power leakage with anywhere from 2X to 5X faster time to design completion.

There were many more examples in the keynote of such advances after an injection of AI into the tools. But here is the thing, and de Geus emphasized this a number of times. The cumulative nature of these advances are not additive, but multiplicative. They will amplify much more than the percents of improvement on many different design vectors might imply. But it is more than that, according to de Geus.

The hand that develops the computer on which EDA is written can help develop the next computer to write better EDA, and so on, de Geus explained at the end of his talk. That circle has brought about exponential achievements. So often we say that success is the sum of our efforts. No, its not. It is the product of our efforts. A single zero, and we all sink. Great collaboration, and we all soar.

Originally posted here:
SysMoore: The Next 10 Years, The Next 1,000X In Performance - The Next Platform

China is winning the AI race due to the US armys sluggish digital transformation, private actors reluctance to work with the state and too many ethical debates stifling innovation, according to the Pentagons former software officer. Market analysts back his diagnosis.

The guys right, Michael Orme, senior analyst at GlobalData and a China specialist, tells Verdict, suggesting that Beijing has leveraged its wealth of data and the surveillance state to gain AI supremacy.

Nicolas Chaillan left the his role as the United States Air Forces first chief software officer in October in protest to the US militarys lethargic digital transformation. He had spent three years working on the Pentagons cybersecurity efforts.

Chaillan is currently listed as the founder and CTO at cybersecurity startup Prevent Breach, which peddles its wares to both federal agencies and commercial organisations, according to his LinkedIn profile.

We have no competing fighting chance against China in 15 to 20 years. Right now, its already a done deal; it is already over in my opinion, Chaillan told theFinancial Times in his first interview after his departure from the Pentagon, adding there was good reason to be angry.

Chaillan argued that Beijing is already set to dominate the world in terms of controlling everything from media narratives to geopolitics, adding that some US government departments digital defences were at kindergarten level.

Chaillan also blamed private actors like Google for the US disadvantage against China, saying the company had been reluctant to collaborate with the US defence department on AI.

By contrast, the cybersecurity founder suggested that Chinas digital defences arent burdened by the same distance between private actors and the government or extensive debates about the ethics of AI.

Chaillan plans to testify before Congress in the coming weeks on the threat of Beijing overtaking the US.

While the FT story suggested that the US has already lost the AI fight to China, Chaillan later clarified on LinkedIn that he didnt believe this was the case, but that the window for the States to act is rapidly closing.

I never said we lost, Chaillan said. I said as it stands and if we dont wake up NOW we have no fighting chance to win against China in 15 years. I also said that theyre leading in AI and Cyber NOW. Not in 10 years as some BS reports mention.

He added: Those are just common sense [facts]. We are competing against [1.5 billion] folks here. Either we are smarter and more agile or we lose. Period.

Explosive as Chaillans comments may seem, analysts have been quick to agree following the publication of the interview.

Xi [Jinping] and the partys top brass had a Sputnik moment in March 2016 when Googles AlphaGo beat the world Go champion, says Orme, who will host a seminar on the Chinese market on Thursday. Since then, its put a heck of a lot of wood behind the AI arrow. China has had the huge advantage, in stark contrast to the West, of its authoritarian system, a generally accepting and patriotic population when it comes to loss of privacy, and a company sector bound by civil-military fusion.

Orme notes that this push in combination with the Chinese surveillance states wealth of data has seen Beijing make up the previous AI gap between China and the US. Orme argues that China will be able to cement its position thanks to projects like the Beijing Academy of Artificial Intelligences recent unveiling of Wu Dao 2.0, an AI multimodal project.

All in all, arguably the West has lost its digital sovereignty in AI to China, Orme says.

Chaillan and Orme are not the only ones who say China is overtaking the US in AI plenty of past reports have highlighted how China has become a global AI leader.

For instance, one of GlobalDatas recent thematic research AI industry scorecards noted that several AI leaders are based in the nation.

In March, the US National Security Commission on Artificial Intelligence (NSCAI) a bipartisan commission of 15 technologists, national security professionals, business executives and academic leaders issued a stark warning in its final report.

America is not prepared to defend or compete in the AI era, the NSCAI report said. This is the tough reality we must face.

Researchers from the University of Electronic Science & Technology of China, and the Leeds School of Business at the University of Colorado issued a similar warning in a recent article in the Harvard Business Review.

They noted that Chinas global share of research papers in the field of AI skyrocketed from 4.26% in 1997 to 27.68% in 2017, surpassing any other country in the world, including the US.

However, those researchers suggested that weak links between Chinese universities, who conduct the majority of the AI research, and private actors could end up stifling innovation.

Similarly, business uncertainty in the country no doubt amplified by Beijings ongoing tech crackdown means that many entrepreneurs are looking for quick, easy wins based on applied AI research that can bring quick money instead of more basic research that promises to have long-lasting impacts.

While Chinese AI developers certainly have many advantages, it is clear they also face several challenges.

Market experts such as Duncan Jones, head of cybersecurity at quantum computer company Cambridge Quantum, believe tech developments in the next few years could see Chinas lead shrink.

No, China has by no means won the AI war, Jones tells Verdict. What we call AI today is but a fraction of what AI will come to mean in the next decade as quantum computers expand the boundaries of possibility. Its far too early to think about winners and losers at this stage.

He adds that despite China having a plethora of data to analyse, it doesnt have the capability to analyse it.

On Chaillans suggestion that ethical considerations have held back US AI innovation, Jones added: Countries such as the USA and the UK have chosen to build and defend a society that might involve different ethical values. The important issue is to think holistically about the structure of justice, democracy and individual rights. We believe societies constantly evolve, so there is no right or wrong amount of ethical considerations. Each society chooses its own balance.

Earlier this month, Verdict reported that China has released a set of ethical guidelines regarding AI, with the focus on enhancing user autonomy and protecting privacy.

Original post:
Yes, the US is losing the AI war to China analysts - Verdict