Merging Man And Machine

When IBM announced on January 9th that it is investing $1 billion in a series of initiatives around Watson, a class of software, services and apps that think, improve by learning, and discover insights and answers to complex questions from massive amounts of Big Data, the company proclaimed a new era of machine-human collaboration, one in which computers and humans do things together to solve problems that neither can do on their own.

Nearly three years after its triumph on the television quiz show Jeopardy!, this advance in cognitive computing, which is due to be discussed on stage at DLD14 by IBM Watson General Manager Manoj Saxena, is expected to have a major impact on key industries, such as healthcare, retail and travel.

Technologies like Watson are needed to help parse what IBM estimates to be 2.5 quintillion bytes of data from sensors, posts to social media sites, research papers, digital pictures and videos posted online, point-of-sale data, transaction records of online purchases, e-mails and mobile phone GPS signals generated by consumers and enterprises every single day.

When put into the right context, this data can be used to figure out who we are, where we are, what we are doing now and what we want to do next, the state of our health and even what we are thinking or feeling — and ideally come up with smart, personalized, actionable insights.

Named after IBM founder Thomas J. Watson, the Watson cognitive computing technology, which was developed in IBM’s Research Labs, aims to do just that. Using natural language processing and analytics, Watson can trawl through the Internet at superhuman speed, digest information it is fed and learn as it goes, processing information akin to the way that people think. And, it can answer complex questions from humans in natural language.

IBM says implementing Watson will change an organization’s ability to analyze, understand and respond to Big Data, helping to do everything from helping doctors to more accurately diagnose patients to improving in-store and online shopping and vacation planning.

IBM is opening up a new headquarters for its Watson business unit in the heart of New York City’s Silicon Alley. In addition to housing its marketing and engineering activities, the building will provide a place for IBM to collaborate with start-ups that are building apps for Watson. IBM said it will invest about $100 million in various start-up companies working on Watson projects.

Watson is just one of a number of advances in technology that can contextualize data about individuals, bringing benefits but also raising concerns, because greater information sharing and more data collection bring risks of information misuse and compromised data security and privacy (see the story on data privacy on page 22).

Transforming Healthcare

At the January 9th New York press conference, which was live streamed, IBM CEO Virginia Rometty promised that Watson “will change the face of healthcare.” And respected physicians agreed.

Two top hospitals, Memorial Sloan-Kettering Cancer Center and the Cleveland Clinic, announced at the press conference that they have had oncologists teaching Watson to recognize different types of cancer and treatments. The goal is to eventually tap Watson’s learning power to deliver better, personalized care to patients.

Watson won’t replace doctors but it can help them improve their diagnoses. Some 400,000 people die in the U.S. alone every year due to preventable medical errors, making it the third leading cause of death, Dr. Jay Katzen, president of clinical solutions at Elsevier, said at the press conference.

While human doctors can’t keep up — there is an average of 5,000 new biomedical articles published every month — Watson can rapidly read the latest findings and apply them to a patient chart. It reads, then adapts its evaluation and potential answers, posting its level of confidence when ranking various diagnoses or treatments.

For example, only a very small percentage of doctors around the world would know about a specific abnormality and be able to prescribe the right treatment with a high level of confidence, as Watson did in an example discussed during the press conference involving a young Asian woman with lung cancer whose genetic make-up required a very particular course of action.

However, because medicine is arguably both an art and a science Watson’s level of confidence in all proposed courses of actions can be quite low, meaning human doctors may still have to make hard choices.

Still, by understanding the context of a patient’s medical history and the latest research Watson is able to act as a decision support tool for medical professionals, helping doctors consider information they might not have and warning against treatments that might have deadly side effects. “I feel strongly that this is going to change the way health care is delivered,” Memorial Sloan-Kettering Physician-in-Chief Jose Baselga said during the press conference.

Watson can also be used by individuals, not just medical professionals. One of the Watson apps that IBM is demoing focuses on preventative care: a 40-year-old “road warrior” is diagnosed with pre-diabetes. Based on what it knows about the businessman and research Watson sends a message to the man’s phone telling him he needs to lose 20 pounds by exercising and controlling his intake to 1,800 calories a day.

Powering Personalized Shopping

“With Watson, you are no longer just creating value for market segments; every interaction is a personalized value creating experience,” says Marc Teerlink, IBM´s Global Strategist for Big Data Analytics, who has worked with clients delivering several Watson-related engagements.

For example, IBM has partnered with Fluid, a U.S. company that specializes in using technology and design to transform shoppers into buyers, to develop an Expert Personal Shopper, powered by Watson, which will allow consumers both online and via kiosks in retail stores to ask product questions in context. Instead of saying, “I want to buy a tent,” a consumer could instead say, “I am taking my family backpacking in Patagonia in the dead of winter and I need gear and other supplies. What should I consider?” Fluid’s Watson-enabled app would explain, in plain language, what kind of gear to buy and would also provide a list of other items that might be needed, drawing from product information, customer loyalty data, sales histories, user reviews, blogs, relevant magazines and publications and travel documents to give users personalized relevant answers.

The same approach could be applied to the travel sector so that people could enter complex queries on line, such as, “Where can I vacation during the month of February that will offer sun, activities for the kids and interesting cultural excursions?” and receive answers that are relevant. Terry Jones, a founder of the travel websites and, said during the press conference that while Internet search engines had become the method of choice for booking travel, they couldn’t yet provide the expert advice about particular destinations and travel activities that an old-fashioned travel agent could, which is why, he sheepishly admitted, he still uses one to book his holidays.

Retailers can also benefit from Watson’s expertise to manage inventory, says Teerlink, by taking both traditional predictive merchandising factors and social sentiment into account (such as how many people liked or mentioned a fur vest worn by Kim Kardashian). Buzz on social networks can influence people’s intention to buy so retailers need this information in order to have the right assortment in every store and make needed adjustments to distribution and delivery planning. “Social data is the new production line,” he says. “It is going to be part of internal prediction, pushing your planning and delivery.”

In a recent study, IBM found that companies that use predictive analytics across multiple channels increased top-line growth five times more than retailers that didn’t (see the chart).

But Teerlink says success is dependent on setting up the system and training Watson on basics about the sector and the company before getting started. And he cautions that cognitive computing “doesn’t have all the answers.” That is where the merger of man and machine comes in.

“If there is a flood in the Philippines and I ask the system what is the impact on electronic components the system wouldn’t know yet. But I could ask the question to the community and my own staff and feed it into the system and it would learn from that,” he says. “I would also make it shareable with vendors, though you have to be careful about who you share it with.”

Data privacy is a concern for both businesses and individuals but a certain amount of openness is a prerequisite to succeeding. “Data sharing is the new black,” says Teerlink. “Facts are better than gut feeling. Today, one in five (business) decisions are not based on facts and that is very scary.”

With the melding of man and machine those odds look set to improve.



Related posts