Archive for February 2011
I’m off to spend a week with mi padre playing golf and running some rapids in Costa Rica.
We arrived at Juan Santamaria aeropuerto last evening and are staying in the “Gringo Gulch” area near downtown San Jose.
I’ve never visited Costa Rica before, so I’m learning as I go. Some factoids from Wikipedia:
- Costa Rica is bordered by Nicaragua to the north (where I haven’t been) and Panama to the east and south (where I have), and the Pacific Ocean to the west and south and the Caribbean Sea to the east.
- Costa Rica has been ranked as one of the happiest places on earth, and consistenly been among the top Latin American countries in the Human Development Index
- Costa Rica is expecting to become the first carbon neutral country by 2021 and is already the “greenest” country in the world.
Economically, Costa Rica has historically centered on coffee, which was first planted here in the early 19th century and first shipped to Europe in 1843.
Bananas later came to rival coffee as the principal Costa Rican export, which were initially shipped primarily back to the U.S.
More recently, turisma (tourism) has become the headliner of the economy here, although high tech is no stranger to Costa Rica: IBM, HP, even Intel, which has a $500+M assembly and test plant here.
So, I’m off to do my part to contribute to the local economy. If you don’t hear much from me in the Turbo blog this week, it’s because I’m chasing little white balls through the rainforest.
But, I’ll be back…Well, maybe.
What’s Next for Watson?
So now that the matches have all aired, and Watson came out the victor, what does it all mean?
If you missed the hullaballoo, this week, IBM’s Watson computer system competed against Jeopardy!’s most successful contestants, Ken Jennings and Brad Rutter, and ended up running away with victory.
But as amazing as the win was, what hasn’t been celebrated nearly enough in this blogger’s opinion is the practical applications of this technology to solve real problems in the world moving forward across a wide range of industries.
Watson was built by a dedicated team of brilliant IBM Research scientists over the past four years, and represents a breakthrough innovation: a machine that rivals a human’s ability to answer questions posed in natural language – quickly, accurately, confidently.
We saw this demonstrated in full force during the Jeopardy! matches this week.
But today begins the next phase of Watson’s evolution.
Calling Doctor Watson
Today, IBM is announcing that doctors from Columbia University Medical Center and the University of Maryland School of Medicine will work to take the same Jeopardy! playing capabilities of Watson and apply them to medicine in an effort to address some of the healthcare industry’s biggest challenges.
IBM also announced a research agreement with Nuance Communications, Inc., to explore, develop, and apply the Watson computing system to healthcare.
Consider these statistics: Primary care physicians spend an average of only 10.7-18.7 minutes face-to-face with each patient per visit. And approximately 81% average 5 hours or less per month – just over an hour per week – reading medical journals.
This results in an estimated 15% of diagnoses being inaccurate or incomplete.
In today’s healthcare environment, where doctors are often working with limited information and little time, the results can be fragmented care and errors that raise costs and threaten quality.
What doctors need is an assistant who can quickly read and understand massive amounts of information and then provide useful suggestions.
Watson’s ability to deal with natural language across a wide collection of diverse information and make it more digestible for humans holds an enormous potential to transform healthcare effectiveness, efficiency and patient outcomes.
Answer: What Is Watson?
Many have probably wondered through the course of the Jeopardy! matches what, exactly is Watson?
Think of Watson as an analytical computing system that specializes in understanding the meaning of natural human language and provides specific answers to questions across a broad domain of knowledge at lightning speeds.
Those domains could span virtually every industry: as noted already in this post, healthcare, but also media/entertainment, financial services, the public sector, transportation, and more.
Watson’s breakthrough comes in its understanding of natural language, not simply language specifically designed and encoded for computers. But, rather, language we humans use to naturally capture and communicate knowledge.
Watson evaluates the equivalent of roughly 200 million pages of content (about 1 million books worth) written in natural human language to find correct responses to the Jeopardy! clues.
The system was named after IBM founder Thomas J. Watson, and is powered by 10 racks of IBM POWER 750 servers and runs on the LINUX operating system. Watson contains 15 terabytes of RAM and 2,880 processor cores, and can operate at 80 teraflops – 80 trillion operations per second.
Yes, you read that correctly. 80 trillion. And yes, I, too, wish I could have that kind of horsepower on my laptop.
Getting A Second Opinion With Watson
As the leading company that helps businesses make sense of data, IBM created Watson to advance our ability to find meaning and knowledge in vast amounts of data.
Watson’s ability to understand the meaning and context of human language, and rapidly process information to find precise answers to questions, holds enormous potential to transform how computers help people accomplish tasks in business and their personal lives.
The fact is, no matter how doctors may try to keep up with the medical literature, it doubles in size every few years and the task of incorporating hundreds of thousands of articles a day into practice and applying them to patient care is difficult and impractical.
Diagnosis, treatment and management of diseases are phenomenally complicated. For any given chronic disease, there are all sorts of nuances. One size doesn’t fit all.
Patients with problems like cancer, diabetes, chronic heart or kidney disease are incredibly complex. But much of that is computable. Watson will be able to look at all that and answer profoundly complex questions by analyzing massive amounts of health data and healthcare knowledge.
Doctors understandably tend to diagnose based on their own specialties or experience. A computer system like Watson can also suggest questions that raise alternatives. For instance, doctors may focus on physical issues and forget to investigate whether symptoms could be caused by depression.
It can narrow among a large group of choices and ultimately help doctors pick the right decision — a very necessary advance in the effective and efficient storage, retrieval, analysis and use of biomedical information to improve health.
IBM calls this Deep DDX – deep differential diagnosis.
Now, lest you have fears that R2D2 will start making the rounds, you’re never going to replace a doctor or a nurse. But, like a good surgical’s assistant, Watson can help inform and advise physicians by creating many hypotheses about a condition then narrowing them down to the one it feels most confident about based on symptoms of the patient, combined with all the information it has received from medical journals, text books, and the like.
Deep QA At Columbia And Maryland
The new research and technology initiative with Nuance will combine IBM’s Deep Question Answering (QA), Natural Language Processing, and Machine Learning capabilities with Nuance’s speech recognition and Clinical Language Understanding (CLU) solutions for the diagnosis and treatment of patients that provide hospitals, physicians, and payers access to critical and timely information.
At Columbia and Maryland School of Medicine, physicians will help identify critical issues in the practice of medicine where the Watson technology may be able to contribute, and to help identify the best way that a technology like Watson could interact with medical practitioners to provide maximum assistance.
Watson’s ability to analyze the meaning and context of human language, and quickly process information to find precise answers can assist decision makers, such as physicians and nurses, unlock important knowledge and facts buried within huge volumes of information, and offer answers they may not have considered to help validate their own ideas or hypotheses.
“Combining our analytics expertise with the experience and technology of Nuance, we can transform the way that healthcare professionals accomplish everyday tasks by enabling them to work smarter and more efficiently,” said Dr. John E. Kelly III, senior vice president and director of IBM Research. “This initiative demonstrates how we plan to apply Watson’s capabilities into new areas, such as healthcare with Nuance.”
For example, a doctor considering a patient’s diagnosis could use Watson’s analytics technology, in conjunction with Nuance’s voice and clinical language understanding solutions, to rapidly consider all the related texts, reference materials, prior cases, and latest knowledge in journals and medical literature to gain evidence from many more potential sources than previously possible.
You don’t want your doctor to guess. You want them to have confidence in their answers.
Built into Watson is the idea of confidence in answering. It’s the doctor’s responsibility to make the diagnoses…Watson would only be there to narrow down the options.
As for Watson’s bedside manners…well, on that one we’ll just have to wait and see.
SPOILER ALERT: Once again, I’m reporting out on the tidings from the Watson v. Jeopardy! “Man v. Machine” matches that has so captured the Web’s attention these past few days. Read no further if you don’t wish to yet know the outcome.
It’s official: Watson smoked the humans. I’m officially adding an exclamation to Watson’s name so he can hereafter be referred to as “Watson!” with an exclamation point.
For those of us inside IBM, again, the surprise is as fresh for us as all of you in the viewing audience. I honestly had no clue who would take it. Jennings and Rutter are formidable competitors.
But Watson! was just too clever and too fast for we mere mortals.
Alex Trebek said it best in the setup of today’s final episode as he explained the three things he’d learned over the last three days about Watson!:
- Watson! is fast. He knows lots of stuff and can really dominate a match.
- Watson! is capable of some weird wagers.
- Toronto is now a U.S. city (see the 2nd day…this comment earned lots of laughs in the audience).
I would add that Watson! isn’t a static entity, but can learn and make smart, deductive predictions both about the success in answering the questions, but to the point of the weird wagers, in “betting” only as much as was needed to take and/or keep his lead.
Watson won the turn and got off to business very quickly, answering a question on the EU, “What is Istanbul?” to get off to a $200 start.
Jennings showed some promise at taking over mid-round, but Watson! came roaring back and ended his combined score at over $77,000, winning the full match.
Jennings and Rutter were both great sports, and all three were able to make significant contributions to the charities of their choice (IBM contributed $1,000,000 to the World Community Grid).
The real question now is, what have we learned?
First, we learned that computers, trained by humans, can outsmart the humans.
But more importantly, we learned we’re now on the precipice of a whole new wave of computing and information technology, one that could very well help mankind arrive at more answers than questions.
In some future posts, I’ll address some of the opportunities and challenges this technology presents to the world, as well as highlight some of those from related media.
Let’s start with this video: Watson! after Jeopardy!
SPOILER ALERT: DO NOT READ if you haven’t watched the second round of Watson and Jeopardy! and are planning on it!
That said, if you had any question as to whether or not Watson could compete on Jeopardy! with world champions Ken Jennings and Brad Rutter, the second round on today’s episode should put your fears to rest.
Watson came out swinging, again, and (mostly) dominated the match.
Through the first 8 questions, Watson rocketed up the scoreboard before wagering $6,435 on a Daily Double (the sum for which earned Watson a laugh from the studio audience, and to which Alex Trebek remarked, “I won’t ask.”)
The question: “The chapels at Pembroke and Emmanuel Colleges were designed by this architect.”
Watson’s answer: “Sir Christopher Wren.”
No easy question, Watson rocks on.
Then, Watson misses a question, but so do Jennings and Rutter and Watson continues control of the game.
At question 10, another Daily Double for Watson: This time it wagers $1,246 and wins that one as well, even though for Watson the answer, “Baghdad,” was only at a 32% confidence.
Watson went on to win the following four questions before Jennings finally took one.
At the end of the day, Watson was at $36,681, Rutter at $5,400, and Jennings at $2,400.
Watson ruled the day.
What wasn’t included in the gameplay was the series of videos interviewing the IBM researchers and executives the implications and opportunities the Watson technology presents to business and the world.
Final, final Jeopardy! and the closing of the match tomorrow. It’s not looking good for the mortals, and the future’s quite bright for the silicon!
I’m no relation to IBM founder Thomas J. Watson, Sr. I joke I’d be on a yacht somewhere if I were.
But I definitely consider myself part of the proud IBM tradition of using technology to solve challenging business problems.
So when our computer (actually, a whole bunch of computers put together along with some transformation Deep Q&A technology developed by IBM Researchers) finally went on air to play the human “Jeopardy!” champions, Brad Rutter and Ken Jennings, I was standing by with baited breath.
DISCLOSURE ALERT: STOP READING NOW IF YOU DON’T WANT TO KNOW WHAT HAPPENED IN THE FIRST EPISODE AND ROUND BETWEEN THE HUMANS AND WATSON
Before the match got underway, Alex Trebek gave an excellent tour of Watson behind the scenes at our Yorktown Research Lab, where these matches were played.
He also set up the match with some color, explaining this was the next “grand challenge in computing” and that we were about to witness what could be an historic competition.
He also joked that Watson would have appreciated the crowd’s applause, but can neither hear nor see, and would receive all questions in a text message at the same time he read them aloud to the human contestants.
Trebek explained Watson would be represented by an avatar, then took us backstage to meet Watson. Watson was set up in two units, he explained, and the first thing you noticed was a lot of “noise,” emanating from two very large refrigerator units to help Watson keep his cool. Literally.
Watson consists of over 2800 POWER750 chips, sitting in some five separate racks on two different units. Linking them all together, you create a deep analytics engine that houses over 15 trillion bytes.
But, Trebek explained in his continuing set up, Watson would have to “stand on his own” and rely on the knowledge that was stored in his memory, and that he couldn’t be connected to the Internet to look anything else up during the match.
“Some of the world’s most brilliant minds have created this most impressive system.”
That was self-evident from the debut of the first match. In a matter of a few minutes, Watson was overpowering Jennings and Rutter, some $5000 to Jenning’s $200 and I think Rutter’s $1,200 or so.
Watson was, in short, smoking the humans, nailing a very tough Double Jeopardy question for $1,000 on the second question.
But, in the second half of the first round, the humans came back in a big way, this time smoking Watson.
It is ON.
However, you won’t be able to fully appreciate what you’re seeing until you learn more about the Deep Q&A technology that was powering Watson.
As for me, I cannot WAIT until the match to pick back up tomorrow (we IBMers also have NO clue who’s going to win).
I want to send my IBM Research colleagues my best wishes — whether Watson wins or not, you have already demonstrated that we’re well into the next frontier of computing, and I’m just glad you’re on our side! LOL
Signed, Todd “Turbo” Watson
The Human (For Now)
P.S. Also human, lead investigator on the Watson/Jeaopardy! initiative, Dr. David Ferucci provides an answer to the question “Why Jeopardy!” in the video below.
Three more days to Watson and “Jeopardy!” IBM productivity across North America will be stunted as laptops go dark and TVs light up to see how Watson fares against the best humans “Jeopardy!” has to offer.
But regardless of the outcome, the technology behind Watson will continue to evolve and be put to productive uses around the world.
And we’re going to have some help.
IBM announced earlier today that eight universities are collaborating with IBM Researchers to advance the Question Answering (QA) technology behind Watson.
Massachusetts Institute of Technology (MIT), University of Texas at Austin, University of Southern California (USC), Rensselaer Polytechnic Institute (RPI), University at Albany (UAlbany), University of Trento (Italy), and University of Massachusetts Amherst join Carnegie Mellon University in working with IBM on the development of a first-of-its-kind open architecture that enables researchers to efficiently collaborate on underlying QA capabilities and then apply them to IBM’s Watson system.
Watson’s QA technology uses breakthrough analytics to understand what is being asked, analyze massive amounts of data, and provide the best answer based on the evidence it finds.
The ongoing research collaborations announced today will help advance Watson’s ability to transform the way businesses and society work and improve all kinds of industries, such as healthcare, banking, government, etc.
“We are glad to be collaborating with such distinguished universities and experts in their respective fields who can contribute to the advancement of QA technologies that are the backbone of the IBM Watson system,” says Dr. David Ferrucci, leader of the IBM Watson project team.
“The success of the Jeopardy! challenge will break barriers associated with computing technology’s ability to process and understand human language, and will have profound effects on science, technology and business.”
I had occasion yesterday afternoon to watch Dr. Ferrucci explain the Watson system via an internal Webcast for a good 90 minutes, and it was replete with specific examples of how he and his team trained and evolved the system over the past four years.
My respect for their Herculean efforts only increased as I realized the complexity of their mission, and the “adaptability” they had to build into the system to get Watson to the level that it could compete with master human players.
Most folks probably don’t realize it, but the Watson system had to “compete” to become an official contestant just like the human champions he’ll be playing against and any other “Jeopardy” player who appears on the show.”
And based on Ferrucci’s deep dive deconstruction of the effort yesterday afternoon, it certainly didn’t happen overnight (It was more like four years).
Once again, I refer you to the IBM/Watson Website, where there are a series of videos in which key participants explain in (mostly) plain English the Deep Q&A technology behind Watson, and the opportunities and implications this technology presents to the world moving forward.
Check the Jeopardy! website to find out when and where you can watch the show in the North American market starting this Monday.
Could you hear the mobile market bombshell Nokia and Microsoft dropped in London earlier today?
Here’s how it was covered on the Nokia Conversations blog. Basically, Nokia’s putting Symbian out to pasture and adopting Windows Phone as its “primary smartphone strategy.”
This will bring Nokia and Microsoft into the same big mobile market bed as they “closely collaborate on development, joint marketing initiatives, and a shared development roadmap.”
It also means that Bing will become the search engine of choice across Nokia’s devices and services. This at a time when Bing’s market share hovered a little over 12% of the U.S., and just under 10% globally.
With the explosion of mobile devices around the globe and Nokia’s vast global market penetration, the Bing deal becomes an attempt to “buy in” as a potential strategic counterweight to Google’s vast and growing mobile search domination.
But considering Nokia’s own 4Q10 press release indicated that YOY market share had dropped from 35% to 31%, they’re going to need a major bounceback.
Then again, nothing surprises me in this industry anymore. Microsoft helped “save” Apple way back when with a $150M cash infusion, so who’s to say Redmond can’t help Nokia get firmly back on the mobile information superhighway.
Though I won’t be running out to replace my own iPhone 4 anytime soon, a device I’ve become perfectly content with (save AT&T’s continued spotty service here in my Austin home office), it will be interesting to see if Redmond can save Helsinki.
A few other key aspects of the deal, FYI, and according to the official press release from Nokia:
- Nokia will help drive and define the future of Windows Phone. Nokia will contribute its expertise on hardware design, language support, and help bring Windows Phone to a larger range of price points, market segments and geographies.
- Nokia Maps will be a core part of Microsoft’s mapping services. For example, Maps would be integrated with Microsoft’s Bing search engine and adCenter advertising platform to form a unique local search and advertising experience.
- Nokia’s extensive operator billing agreements will make it easier for consumers to purchase Nokia Windows Phone services in countries where credit-card use is low.
- Microsoft development tools will be used to create applications to run on Nokia Windows Phones, allowing developers to leverage the ecosystem’s global reach.
- Microsoft will continue to invest in the development of Windows Phone and cloud services
- Nokia’s content and application store will be integrated with Microsoft Marketplace