Connect with us

Technology

Making New Drugs With a Dose of Artificial Intelligence

Published

on


SAN FRANCISCO — You can think of it as a World Cup of biochemical research.

Every two years, hundreds of scientists enter a global competition. Tackling a biological puzzle they call “the protein folding problem,” they try to predict the three-dimensional shape of proteins in the human body. No one knows how to solve the problem. Even the winners only chip away at it. But a solution could streamline the way scientists create new medicines and fight disease.

Mohammed AlQuraishi, a biologist who has dedicated his career to this kind of research, flew in early December to Cancun, Mexico, where academics were gathering to discuss the results of the latest contest. As he checked into his hotel, a five-star resort on the Caribbean, he was consumed by melancholy.

The contest, the Critical Assessment of Structure Prediction, was not won by academics. It was won by DeepMind, the artificial intelligence lab owned by Google’s parent company.

“I was surprised and deflated,” said Dr. AlQuraishi, a researcher at Harvard Medical School. “They were way out in front of everyone else.”

DeepMind specializes in “deep learning,” a type of artificial intelligence that is rapidly changing drug discovery science. A growing number of companies are applying similar methods to other parts of the long, enormously complex process that produces new medicines. These A.I. techniques can speed up many aspects of drug discovery and, in some cases, perform tasks typically handled by scientists.

“It is not that machines are going to replace chemists,” said Derek Lowe, a longtime drug discovery researcher and the author of In the Pipeline, a widely read blog dedicated to drug discovery. “It’s that the chemists who use machines will replace those that don’t.”

After the conference in Cancun, Dr. AlQuraishi described his experience in a blog post. The melancholy he felt after losing to DeepMind gave way to what he called “a more rational assessment of the value of scientific progress.”

But he strongly criticized big pharmaceutical companies like Merck and Novartis, as well as his academic community, for not keeping pace.

“The smartest and most ambitious researchers wanting to work on protein structure will look to DeepMind for opportunities instead of Merck or Novartis,” he wrote. “This fact should send chills down the spines of pharma executives, but it won’t, because they’re clueless, rudderless, and asleep at the helm.”

The big pharma companies see the situation differently. Though Merck is not exploring protein folding because its researchers believe its potential impact would be years away, it is applying deep learning to other aspects of its drug discovery process.

“We have to connect so many other dots,” said Juan Alvarez, associate vice president of computational and structural chemistry at Merck.

In the spring of 2016, after making headlines with A.I. systems that played complex games like the ancient board game Go, DeepMind researchers were looking for new challenges. So they held a “hackathon” at company headquarters in London.

Working with two other computer scientists, the DeepMind researcher Rich Evans homed in on protein folding. They found a game that simulated this scientific task. They built a system that learned to play the game on its own, and the results were promising enough for DeepMind to greenlight a full-time research project.

The protein folding problem asks a straightforward question: Can you predict the physical structure of a protein — its shape in three dimensions?

If scientists can predict a protein’s shape, they can better determine how other molecules will “bind” to it — attach to it, physically — and that is one way drugs are developed. A drug binds to particular proteins in your body and changes their behavior.

In the latest contest, DeepMind made these predictions using “neural networks,” complex mathematical systems that can learn tasks by analyzing vast amounts of data. By analyzing thousands of proteins, a neural network can learn to predict the shape of others.

This is the same deep learning technology that recognizes faces in the photos you post to Facebook. Over the past decade, the technology has reinvented a wide range of internet services, consumer products, robotic devices and other areas of scientific research.

Many of the academics who competed used methods that were similar to what DeepMind was doing. But DeepMind won the competition by a sizable margin — it improved the prediction accuracy nearly twice as much as experts expected from the contest winner.

DeepMind’s victory showed how the future of biochemical research will increasingly be driven by machines and the people who oversee those machines.

This kind of A.I. research benefits from enormous amounts of computing power, and DeepMind can lean on the massive computer data centers that underpin Google. The lab also employs many of the world’s top A.I. researchers, who know how to get the most out of this hardware.

“It allows us to be much more creative, to try many more ideas, often in parallel,” said Demis Hassabis, the chief executive and a co-founder of DeepMind, which Google acquired for a reported $650 million in 2014.

Universities and big pharmaceutical companies are unlikely to match those resources. But thanks to cloud computing services offered by Google and other tech giants, the price of computing power continues to drop. Dr. AlQuraishi urged the life-sciences community to shift more attention toward the kind of A.I. work practiced by DeepMind.

Some researchers are already moving in that direction. Many start-ups, like Atomwise in San Francisco and Recursion in Salt Lake City, are using the same artificial intelligence techniques to accelerate other aspects of drug discovery. Recursion, for instance, uses neural networks and other methods to analyze images of cells and learn how new drugs affect these cells.

The big pharma companies are also beginning to explore these methods, sometimes in partnership with start-ups.

“Everyone is trending up in this area,” said Jeremy Jenkins, the head of data science for chemical biology and therapeutics at Novartis. “It is like turning a big ship, and I think these methods will eventually scale to the size of our entire company.”

Mr. Hassabis said DeepMind was committed to solving the protein folding problem. But many experts said that even if it was solved, more work was needed before doctors and patients benefited in any practical way.

“This is a first step,” said David Baker, the director of the Institute for Protein Design at the University of Washington. “There are so many other steps still to go.”

As they work to better understand the proteins in the body, for instance, scientists must also create new proteins that can serve as drug candidates. Dr. Baker now believes that creating proteins is more important to drug discovery than the “folding” methods being explored, and this task, he said, is not as well suited to DeepMind-style A.I.

DeepMind researchers focus on games and contests because they can show a clear improvement in artificial intelligence. But it is not clear how that approach translates to many tasks.

“Because of the complexity of drug discovery, we need a wide variety of tools,” Dr. Alvarez said. “There is no one-size-fits-all answer.”



Source link

قالب وردپرس

Technology

Facebook’s Open Compute Project hits over $2.5 billion in revenues

Published

on

By


The Open Compute Project began its life at Facebook as a revolutionary idea to do for data center hardware what Linux and open source software did for the software market. In other words, the OCP comes up with cutting-edge, super-efficient designs that any company can use to build their own hardware.

And the OCP has succeeded, by most reasonable metrics. The project has created a fanatical following among data center engineers, and has led to the creation of products in 10 categories, including networking, servers, and storage.

And in terms of dollars and cents: On Thursday, the preliminary results of a new market assessment report commissioned by OCP was released. That report finds that companies spent more than $2.5 billion on OCP-designed products in, up from $1.16 billion year the before.

Read: AT&T signed an ‘8-digit’ deal that isn’t good news for VMware, Cisco, or Huawei — but could be great for Google Cloud

And this report doesn’t actually reveal the true amount being spent on OCP gear.

It deliberately hides what the project’s board member companies are spending on their OCP equipment, which includes Facebook, Goldman Sachs, Intel, Microsoft and Rackspace. Those companies run enormous data centers and buy a lot of data center equipment, meaning the real figure is likely higher.

The reason the board members are excluded is to try and show that the project is having an impact beyond the handful of companies in leadership roles — although it’s a bit coy of the organization to keep mum on how much money those companies pour into the ecosystem.

Even so, the commissioned report makes a fair case that OCP is creating a multi-billion market.

Excluding the purchases of board members, OCP products account for nearly 1% of the total data center market, which it pegs at $127 billion, the report says.

Interestingly, the report also finds that the overall data center equipment market is shrinking, from $137 billion in 2017 to $127 billion in 2018. Companies across the board are reducing their use of private data centers, as their use of the cloud increases. And OCP includes many of the big cloud providers that are taking those workloads, including Microsoft, Google and Rackspace.

Simply put, that means that OCP has been eating the data center market in a measurable way.

Engineers love it

OCP’s goal is to take the power out of the hands of traditional server and networking vendors like Hewlett Packard Enterprise, Dell, or Cisco, and put it into the hands of the companies who buy and use that hardware.

While all three of those companies have joined the project, OCP members design their own servers, storage, and networking gear, making them cost less and perform faster than traditional commercial alternatives. Then, they share their designs for free. Anyone can modify those designs for their own use, or share them with the group.

Attendees at the OCP Summit conference
OCP

Engineers love it. They get to freely collaborate with other top engineers trying to solve the same problems without worrying about protecting intellectual property or trade secrets.

Contract manufactures are available to build the gear, too, to make it easier for even smaller companies to take advantage of OCP gear.

OCP has also become such a big thing that a growing list of vendors, including HPE and Dell, also make commercial products that match OCP specifications. So OCP-designed products can be bought off-the-shelf. They don’t have to be custom-ordered, lowering the bar to entry.

Next up: the telecom industry

With a loyal following of data center engineers, OCP and Facebook have moved on to a related industry: telecom equipment.

Through OCP, telecom providers like AT&T and Deutsche Telekom are working on open source designs for routers and the other equipment that run their networks. This is gear that would challenge networking giants including Cisco and Juniper.

A few years ago, Facebook also launched a telecom-specific organization called the Telecom Infra Project. It is working on projects like open source telecom radio transmitters. This is gear that would take on the likes of Ericsson, Nokia and Huawei at this especially critical time, when telcos are upgrading their networks to 5G.

Meanwhile, the telecom industry has also decided that it wanted to lead its own open-source hardware project, away from Facebook.

A project called the O-RAN Alliance has gained steam, and includes a who’s who of the major telecom companies worldwide. This includes AT&T, T-Mobile, Verizon, Sprint, SoftBank, SK Telecom, Telefónica, and others.

The industry scuttlebutt is that the two groups, TIP and O-RAN, are going to announce some sort of collaboration next week at Mobile World Congress so they don’t duplicate efforts as they work to to upend the global telecom equipment industry.

Read: Bill Gates warns of the dangers of cow farts — and the world should take his words seriously

Amy Wheelus, AT&T VP of Network Cloud & Infrastructure heads Airship
YouTube/TelecomTV

OCP’s market research report doesn’t shed much light on how much money the telecos might shift to these new open source creations.

But it does show that telecom companies are one of the major users of OCP gear — including servers, storage and OCP’s optical networking equipment.

Meanwhile AT&T has taken open source even further. It’s leading a project called Airship to share software that it’s building to run and manage its 5G network. This software can be used for lots of other data center needs at all sorts of other companies.

The radical idea that launched OCP is turning into a full-fledged hardware industry coup.



Source link

قالب وردپرس

Continue Reading

Technology

Roku CFO Steve Louden says it’s in a good position in ad-based video

Published

on

By


Roku isn’t worried about Amazon or anyone else horning in on its cash cow.

The electronics maker has transformed itself in recent years into an advertising business, thanks in no small part to the Roku Channel, an advertising-supported free streaming video service that’s available on Roku devices and through the web. Last month, though, Amazon launched a rival service called Freedive from its IMDb unit that threatens to steal viewers and ad dollars from Roku’s offering.

But that’s not how Steve Louden, Roku’s chief financial officer, sees it. Amazon’s entry — along with similar services from YouTube, Vudu, and others — just serve as “validation” for the Roku Channel and the ad-supported streaming business in general.

“We’re strong supporters of ad-supported content,” Louden told Business Insider in an interview on Thursday, just after the company reported its fourth-quarter results.

Read this:Amazon’s got its eyes set on yet another market — and one high-flying upstart should be worried

Roku topped analyst expectations as revenue from its platform business — which includes its advertising sales — jumped 77% from the holiday period of 2017.

Roku is in “a strong position”

The streaming video company is in a better position than many of its rivals to capitalize on ad-supported video market, Louden said. Its control of not just a streaming channel, but a streaming media platform — through its Roku streaming boxes and smart televisions that run its operating system — gives it important data on users’ viewing habits that competitors don’t have, he said. Through its platform, Roku also has the ability to steer viewers to the Roku Channel and other places that run its video ads.

“That puts us in a strong position,” he said.

Amazon too has its own platform in the form of its Amazon Fire TV devices, and it has plenty of data on viewing habits through that, its Amazon Fire tablets, and its Prime Video service. But Louden seemed unconcerned, suggesting that Amazon and many of Roku’s other competitors can’t fully match up with it. Roku can offer advertisers both the data they need to target their ads and a large viewership for them.

“That’s where a lot of folks have gaps,” he said.

Here’s what Roku reported and how it compared with Wall Street’s expectations:

  • Fourth-quarter (Q4) revenue: $275.7 million. Analysts had forecast $262.4 million.
  • Q4 earnings per share (EPS): 5 cents. Wall Street was expecting 3 cents a share.
  • First-quarter (Q1) revenue (company guidance): $185 million to $190 million. Analysts had projected $188.8 million.
  • Q1 EPS (guidance): Roku forecast that it will lose $28 million to $32 million, which works out to a per-share loss of 23 to 26 cents, assuming its share count stays stable. Wall Street was forecasting a loss of 12 cents a share.
  • 2019 full-year revenue (guidance): $1 billion to $1.025 billion. Analysts had forecast for $985.4 million.
  • 2019 EPS (guidance): The company projected a loss of between $80 million and $90 million, which is about 65 cents to 73 cents a share, assuming its share count remains the same. Analysts had predicted a full-year loss of 23 cents a share.

Roku’s stock jumped $2.72, or 5%, to $54.20 in after-hours trading. Its shares closed regular trading off $2.16, or 4%, to $51.48.



Source link

قالب وردپرس

Continue Reading

Technology

The Trade Desk’s Q4 2018 earnings beat expectations

Published

on

By


The Trade Desk continues to be an outlier among ad-tech companies that struggling to grow ad revenue as more of those dollars go to Facebook and Google.

The company reported $160.6 million in fourth-quarter 2018 revenue on Thursday, primarily boosted by growth in programmatic ad dollars flowing to connected TV devices and audio.

The programmatic advertising firm reported a total of $477 million in revenue during 2018, up 55% from $308.2 million in 2017. The Trade Desk’s technology plugs into agency trading desks to power programmatic advertising.

In 2019, The Trade Desk said that it expects to grow faster than the rest of the programmatic industry, making at least $637 million with gross spending on its platform hitting at least $3.2 billion, said Jeff Green, The Trade Desk’s CEO, during the earnings call.

Programmatic firms are making connected TV gains

The Trade Desk saw the biggest growth from connected TV, where spending grew 525% year-over-year. Mobile spend jumped 69%. while programmatic audio spending grew 230%.

During the fourth-quarter, more than 160 advertisers spent more than $100,000 each on connected TV advertising, Green said. In 2018, the company’s inventory for streaming TV ads grew sixfold, with the bulk of new inventory coming from networks like NBCUniversal, A+E Networks and CBS that are building their own streaming services.

Read more: Ad-tech companies and networks are pinning hopes on streaming TV, but OTT is full of headaches for marketers

He added that inventory is also coming from digital players like Hulu, which works with The Trade Desk to power programmatic advertising.

But streaming TV ads are significantly more expensive with higher cost-per-impressions (or CPM) prices than display ads. Over time, prices will come down as more premium content becomes available, Green said.

“I don’t think it will have any big, long-term effect on our fee structure because we add so much more value by bringing data to the table,” he said. “Time will tell there but I think we’re in a really strong position.”

This week, big brands like McDonald’s and AT&T pulled their YouTube ads after it was revealed that ads ran alongside videos with inappropriate comments. Asked about what the pushback against YouTube means for The Trade Desk, Green said that he expects to see a short-term increase in spending from big advertisers over the coming weeks.

“There’s a bunch of dollars that need to find a new home,” he said. “I do think it represents an opportunity for us, but I think it’s hard for all those advertisers to move away from YouTube.”

China holds a lot of potential

The Trade Desk’s move into China was another big topic on the earnings call. The Trade Desk has long eyed Asia as a source for growth and analysts repeatedly asked Green for details on the company’s plans, particularly in China. According to Green, 86% of the firm’s revenue comes from the US, with the goal to get two-thirds of revenue from international markets.

“The fastest-growing and largest middle class in the history of the world is emerging here in Asia, and global brands want to reach these new consumers,” Green said.

Specifically, Green mentioned Alibaba, Baidu and Tencent as critical media partners in Asia. However, the Chinese market is notoriously difficult for marketers to crack. Green emphasized that the country is a “long-term investment.”

Because the Chinese companies have been slower to ramp up advertising, Green said that they have a benefit from learning from Facebook, Google and Amazon’s measurement mistakes and walled gardens.

“There’s actually clearer lines with Baidu, Alibaba and Tencent than there is with Google, Amazon and Facebook, which makes it much easier to have conversations about activating data,” he said. “I don’t think we’re going to have the same debates and evolution that we had in the rest of the world.”



Source link

قالب وردپرس

Continue Reading

Chat

Trending