IBM has spent more than $4 billion buying a handful of companies with vast stores of medical data for Watson AI.
Adam Rifkin stashed this in Machine Learning
IBM Is Counting on Its Bet on Watson, and Paying Big Money for It - The New York Times:
IBM has spent more than $4 billion buying a handful of companies with vast stores of medical data. “A.I. machines are only as smart as the data you give them,” said John E. Kelly, who oversees IBM’s research labs and the Watson business.
IBM has acquired industry expertise and proprietary data sources, in addition to supplying A.I. platform tools and technology. “IBM has taken a very different approach from Google, Amazon and Microsoft,” said Judith Hurwitz, an independent analyst. “It’s leveraging data and the knowledge of experts.”
IBM first focused on health care, and that business now accounts for two-thirds of the Watson unit’s employment. Three years ago, IBM experts began working with leading medical centers. And it has spent more than $4 billion buying a handful of companies with vast stores of medical data like billing records, patient histories, and X-ray and M.R.I. images. “A.I. machines are only as smart as the data you give them,” Mr. Kelly noted.
IBM's Watson was tested on 1,000 cancer diagnoses made by human experts at the University of North Carolina School of Medicine. In 99 percent of them, Watson recommended the same treatment as the oncologists. In 30 percent of the cases, Watson found a treatment option the human doctors missed. Some treatments were based on research papers that the doctors had not read. More than 160,000 cancer research papers are published a year.
An estimated 14 million Americans are living with cancer.
AI is a big market, getting bigger.
The market — defined as A.I.-related hardware, software and services — will surge from $8 billion this year to $47 billion by 2020, predicts IDC, a research firm.
IBM does not report financial results for Watson separately. But the securities research arm of the Swiss bank UBS estimates that Watson may generate $500 million in revenue this year and could grow rapidly in the years ahead, possibly hitting nearly $6 billion by 2020 and almost $17 billion by 2022.
By 2020, IDC predicts, 60 percent of the A.I. applications will run on the platform of four companies: Amazon, Google, Microsoft and IBM.
Winner take all?
THE A.I. RACE
Tech firms are rushing to become the leading platform in the emerging artificial intelligence marketplace. Here are six companies to watch:
- AMAZON. The front-runner in cloud computing is investing in A.I. expertise — to improve its own commerce business and to offer some A.I. to outside developers.
- MICROSOFT. The builder of the winning platform of the PC era is strong in cloud computing and A.I. research. It has open-sourced its “deep learning” A.I. technology to attract developers.
- GOOGLE. The biggest user of A.I.-related technology in-house, from search to self-driving cars. It has open-sourced its machine-learning A.I., TensorFlow, and young developers seem to love its tools.
- IBM. Watson is impressive, and Big Blue has deep ties in the corporate world and lots of industry expertise to apply A.I. technology.
- SALESFORCE. This pioneer in cloud computing for the corporate market just announced its A.I. platform, Einstein. The plan, for now, is to add A.I. smarts to sales and customer service apps.
- BAIDU. “China's Google” is investing heavily in A.I. in a drive led by Andrew Ng, a deep-learning star formerly at Google.
AI as a service!
They point to a new Watson offering in genomics as a prime example of the company’s strategy. IBM is collaborating with Quest Diagnostics, the medical laboratory company, to offer gene sequencing and Watson diagnostic analysis, as a cloud service, to oncologists treating cancer patients.
There is simply too much information out there for one person to know.
1400+ Reddit comments:
Joseph Smarr on why cleaning up data is hard: