Bits Blog: I.B.M. Research Points to Circuits That Mimic the Brain’s Design

A nanofluidic circuit would operate by passing ionic fluid, shown in green, through conduits fabricated on top of a planar oxide surface, shown in orange.

I.B.M. scientists said Thursday that they had developed a fluidic electronic system that mimics the circuits in the human brain and potentially offers a new direction for ultra-low-power microelectronics and artificial intelligence.

A group of researchers at the company’s Almaden Research Center in San Jose, Calif., reported in the journal Science that they had pioneered a novel mechanism for transforming an insulating material into a metallic conductor by placing it in contact with a charged fluid. In contrast to conventional semiconductors, which use electric currents to switch materials between insulating and conducting states, the new method uses what the researchers describe as “ionic currents” — mobile charged atoms rather than electrons — as a switching mechanism.

“I’m particularly excited by our findings,” said Stuart Parkin, a physicist and I.B.M. Fellow, “because a lot of how the brain operates is by the flow of ions and ion channels. In some sense what we want to do is mimic those components of the brain.”

While the individual components of the brain work far more slowly than modern microelectronic transistors, the brain’s circuits are arranged in three dimensions and operate in parallel. That allows the brain to do complex computing using only a fraction of the energy of today’s computers.

The I.B.M. researchers hope that their approach could be used to build more brain-like computers.

The advantage of the new method is that it is both nonvolatile — it requires only a small amount of electricity to change the materials from one state to another, and they then remain in that state — and is potentially reversible, meaning that it could be used to build a device like a transistor.

The researchers noted that while the switching speed of the new materials might never match the raw speed of today’s transistors, their biological-like qualities might make them appropriate for building a new generation of sensors or memories.

Although the initial I.B.M. results are based on simply exposing oxidized materials to fluids, the researchers said that if systems were built upon the new mechanism, they could exploit fields that are known as nano- or microfluidics. These technologies use tiny channels and pipes to control and mix fluids for a variety of industrial and scientific applications.

The next step for the I.B.M. research team would be to make “fluidic” circuits in which it would be possible to move the charged fluids over surfaces to change their properties, much as a conventional microelectronic semiconductor is switched “on” and “off.”

“We could form or disrupt connections just in the same way a synaptic connection in the brain could be remade, or the strength of that connection could be adjusted,” Dr. Parkin said.

Analysts said I.B.M.’s announcement was likely to touch off broader interest in the field within the scientific community.

“This could have applications from fluidics to nonvolatile electronics to chips that are immune from radiation,” said Richard Doherty, an analyst at the Envisioneering Group, a technology research firm.

Dr. Parkin said the I.B.M. scientists were still considering which direction to pursue with their new materials. “Probably initially we’ll build a small memory array or something like that,” he said.

The I.B.M. research is in a field known as correlated electron systems, which explores a wide range of materials that exhibit unusual electronic or magnetic behavior.

Huawei to Open Research Center in Finland

PARIS — Huawei Technologies, a Chinese maker of telecommunications equipment, said on Monday that it planned to open a research and development center in Helsinki next year, accelerating its investments in Europe, where its business is expanding rapidly.

The move illustrates a trans-Atlantic difference in attitudes toward Huawei. The company has been largely shut out of the United States market for network gear because of Congressional concerns about possible security threats — fears the company insists are unfounded.

While Huawei has faced difficulties in some European markets, like France, it has done better elsewhere. Huawei employs more than 7,000 people in the region, and it says that total could double in the next three to five years. Huawei already has a research center in Italy and is studying the possibility of opening one in Spain. It also recently announced a $2 billion investment in Britain.

The planned center in Helsinki, involving an investment of 70 million euros, or about $90 million, will work on smartphone development, including features like user interfaces and power management, the company said. When the center opens next year, it will employ 30 people, but this could grow to 100 over the next five years, the company said.

The announcement is a plus for the Finnish technology industry, which has been suffering from the woes at Nokia. The company was once the world’s biggest cellphone maker, but its market share has fallen sharply in recent years.

“The open and innovative environment in Finland,” Huawei said, “is an ideal place for Huawei to strengthen our global R.& D. capabilities for devices, creating opportunities for both Huawei and the Finnish telecommunications industry.”

Huawei has been known mostly for its network equipment, but the company is pushing to make a name for itself with its handsets.

Mobile devices accounted for 22 percent of its revenue last year, an increase of 37 percent. That compares with growth of 12 percent for the overall business.

Bits Blog: Research Firm Says Windows 8 Had a Rocky Start

Elaine Thompson/Associated Press Representatives from personal computer companies whose devices run Microsoft 8.

For much of the week, Microsoft has been trumpeting the strong start of Windows 8. On Tuesday, Tami Reller, chief financial officer and chief marketing officer for Windows, told investors that Microsoft had sold 40 million licenses to the operating system during its first month on the market. Steven A. Ballmer, Microsoft’s chief executive, promoted that figure at the company’s annual shareholder meeting on Wednesday.

But on Thursday, NPD, the retail sales tracking firm, published data that painted a darker picture of the Windows 8 introduction.

Unit sales of Windows PCs in retail stores in the United States fell 21 percent in the four-week period spanning Oct. 21 to Nov. 17, compared to the same period the previous year, according to the firm.

NPD said sales of Windows 8 tablets had been “almost nonexistent,” accounting for less than 1 percent of all Windows 8 device sales. NPD collects sales data in weekly increments, so the week of Oct. 21 was the first in its system to include sales of Windows 8, which went on sale on Oct. 26.

The figures suggest that Windows 8 did nothing to arrest the downward trajectory of the PC business, much less lead to a rebound in a market that has been struggling for some time. “It hasn’t made the market any worse, but it hasn’t stimulated things either,” Stephen Baker, an analyst at NPD, said in an interview. “It hasn’t provided the impetus to sales everybody hoped for.”

There are a few things to keep in mind about what NPD does and does not count in its figures. The firm tracks sales to customers in most, but not all, American retail stores. In other words, it tracks actual consumer demand for products — in this case, Windows PCs.

Microsoft’s 40 million figure, in contrast, represents copies of Windows that Microsoft sells to all of its customers. That includes some consumers but more often it reflects sales to the hardware makers that install Windows on their machines, some of which have not yet been bought by consumers.

Microsoft’s figure also includes sales to people who are buying the new operating system to upgrade existing PCs, along with sales to business customers that don’t happen at cash registers in stores, which NPD doesn’t reflect. Typically, though, new versions of Windows are adopted most quickly by consumers, so it’s not clear that the inclusion of the business market would show a positive direction for the PC business.

It’s possible also that Black Friday, which was not included in NPD’s figures, provided a jolt to PC sales.

The Windows 8 debut looks like it had much less of a positive impact on PC sales than did its predecessor, Windows 7, which went on sale to the general public on Oct. 22, 2009.

At the time, NPD said that unit sales of Windows PCs rose 49 percent during the first week Windows 7 was on sale, compared to the same period the previous year. Mr. Baker said he wasn’t able to provide sales data for the first four weeks of Windows 7’s availability for a more complete comparison to Windows 8.

The PC business in 2007 had much stronger unit sales than it has now, in large part because of a boom in the low-cost laptops known as netbooks. Fast forward to 2012, and sales of netbooks have nearly vanished, replaced by surging sales of the iPad and other tablets.

Unboxed: Making the Case for a Government Hand in Research

Ambitious, sure, but Duolingo recently attracted $15 million of venture capital. The investors are betting on Mr. von Ahn, his idea and his growing team of 18 engineers, language experts and Web designers.

Mr. von Ahn, 33, personifies some of the essential ingredients of America’s innovation culture, when it works well. An immigrant from Guatemala, he has intelligence and entrepreneurial energy to spare. And he has received a helping hand from the federal government. Duolingo began as a university research project financed by the National Science Foundation.

That pattern has been repeated countless times over the years. Government support plays a vital role in incubating new ideas that are harvested by the private sector, sometimes many years later, creating companies and jobs. A report published this year by the National Research Council, a government advisory group, looked at eight computing technologies, including digital communications, databases, computer architectures and artificial intelligence, tracing government-financed research to commercialization. It calculated the portion of revenue at 30 well-known corporations that could be traced back to the seed research backed by government agencies. The total was nearly $500 billion a year.

“If you take any major information technology company today, from Google to Intel to Qualcomm to Apple to Microsoft and beyond, you can trace the core technologies to the rich synergy between federally funded universities and industry research and development,” says Peter Lee, a corporate vice president of Microsoft Research. Dr. Lee headed the National Research Council committee that produced the report, titled “Continuing Innovation in Information Technology.”

The long-term importance of government-supported research may loom small in the current debate over how to reduce the federal deficit. But it is an economic issue worth keeping in mind, and one that points to the kinds of tough choices and trade-offs facing policy makers.

The Budget Control Act, which is scheduled to take effect next January unless Congress shifts course, calls for across-the-board cuts in discretionary spending — for programs other than entitlements like Social Security and Medicare. A new study by the American Association for the Advancement of Science estimates that federal spending on research and development would be trimmed by more than $12 billion in 2013. The National Science Foundation, which finances most government-supported computer science research at universities, would have its budget cut by more than $450 million.

Last week, Senate leaders were trying to negotiate a deficit-reduction deal that would avoid the automatic cuts, but programs like those that finance research are likely to receive rigorous scrutiny from budget-cutters for years into the future. Already, the Advancement of Science report says, government spending on research and development has declined by 10 percent since 2010, when adjusted for inflation.

YET why should government support for scientific research and technology development be spared from the belt-tightening? Unless society benefits inordinately from such spending, there is no case for special treatment. In a new book, “Innovation Economics: The Race for Global Advantage” (Yale University Press), Robert D. Atkinson and Stephen J. Ezell forcefully present the argument for the exceptional role that science and technology play in the economy.

Cutting funding for research and development, Mr. Atkinson said in an interview, is “completely shortsighted.” Spending on science and technology, he said, is an investment that produces a larger economy in the future — generating wealth, jobs and tax revenue. Besides, he said, the bill is not very high in the United States. America ranks 22nd among 30 nations in university funding and research and development funding as a share of the economy’s gross domestic product.

In their book, Mr. Atkinson, president of the Information Technology and Innovation Foundation, a nonprofit policy research group, and Mr. Ezell, a senior analyst at the foundation, define innovation as not only the generation of new ideas but also as their adoption in new products, processes, services and organizational models. In their view, the goal of policy should be to invest in and nurture the development of the innovation pipeline, from basic science to commercialization.

That would call for a more hands-on role for government than is embraced by the mainstream of economic thought, certainly in the United States. The consensus of most economists is that basic science is a “public good,” with the benefits widely shared by society, and thus a worthwhile recipient of government financing. But technology — the application of science to real-world problems — is regarded as a “private good,” with its development best left to the marketplace.

Mr. Atkinson, whose Ph.D. is in city and regional planning, says he is presenting the case for loosening the grip of “neoclassical economists” on policy. He has held economic and technology policy jobs in Rhode Island and for Congress, and has served in advisory groups in the administrations of President Obama, Bill Clinton and George W. Bush. Mr. Atkinson’s nonprofit policy research organization receives financial support from groups including the Alfred P. Sloan Foundation and the Ewing Marion Kauffman Foundation, and from corporations including Intel and I.B.M.

A linchpin of innovation policy, according to Mr. Atkinson, is collaboration between government and industry. As a prime example, he points to Germany and its network of 60 Fraunhofer Institutes, financed 70 percent by business and 30 percent by federal and state government. The institutes, he says, perform applied research intended to translate promising technologies, from polymer research to nanotechnology, into products. These tech-transfer clusters, Mr. Atkinson says, are an important reason for the strength of Germany’s manufacturing sector, even though wages for its factory workers are 40 percent higher than those for American workers.

Taking a page from the German model, the Obama administration announced this year plans for up to 15 manufacturing innovation institutes, public-private collaborations called the National Network of Manufacturing Innovation. The first will be in Youngstown, Ohio, specializing in custom manufacturing using 3-D printing technology. Mr. Atkinson says that while this is a good step, what is needed is a long-term commitment. He noted that Germany’s Fraunhofer initiative began nearly four decades ago, and has grown steadily.

Mr. Atkinson’s focus on research investment, and on the pathways that bring ideas into the marketplace, is gaining increasing attention in economics, according to David B. Audretsch, an economist at Indiana University. “Research and development unequivocally pays off economically,” Mr. Audretsch says. “But the biggest payoff is in ecosystems that take the innovative inputs and make them commercial outputs — products.”

MR. VON AHN’S start-up is a new addition to the growing technology cluster around Pittsburgh, with Carnegie Mellon, a major research university, as its hub. Duolingo combines crowdsourcing and computing, language learning and online translation. People learn a language at no cost, while their lessons feed into Duolingo’s fast-expanding translation database, which is sorted and scored for accuracy with smart software.

The crowdsourcing principle is similar to that used by Mr. von Ahn’s previous company, reCaptcha. Its service employs the widely used Web security feature in which users identify and type words, scanned from old books and newpapers, to gain access to a site or service — in the process helping to accurately digitize those old texts.

But Duolingo is far more ambitious and complex. The exploratory research was supported by a five-year grant from the National Science Foundation, about $120,000 a year, used mostly to pay the tuition and living expenses for a graduate student. Two years later, Mr. von Ahn and his small team had made enough progress to have a working prototype, start the company and attract private investors.

“The grants are really helpful when you’re working on something that is scientifically complex,” Mr. von Ahn says, “when you’re in the early year, or two or more, when you have no idea if it will work.”

Mr. von Ahn, joining other scientists, has made a few trips to Washington to speak to members of Congress and their staffs. His message is this: “Don’t eat your seed corn. It may seem like an easy thing to cut now. But years later, you will regret that you did not invest a tiny portion of your federal budget in research, in the future.”