IBM Watson Vs. Amazon: Machine Learning Systems Presage the Future

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

IBM has traveled a long path to make artificial intelligence (AI) more accessible. Its Big Blue supercomputer famously beat world chess champion Gary Kasparov and Watson captured the public’s attention when it beat the reigning Jeopardy champion Ken Jennings.

And a recent and amusing ad for Watson features Bob Dylan with Watson telling the iconic folksinger that he can read 800 million pages a second. Intones Dylan, dryly: “That’s fast.”

Such high profile stunts are great publicity, but enterprises are more interested in practical applications than winning games or analyzing all of Dylan’s song lyrics (which Watson did). Fortunately IBM has delivered, evolving Watson into a service used across a variety of industries from medicine to sports. Watson has its own developer ecosystem and a cloud-based platform called IBM Bluemix that gives developers the opportunity to tap into Watson’s capabilities to create new “intelligent” apps.

The Bluemix community consists of 77,000 developers who are prototyping and building what IBM’s calls “cloud-based cognitive computing applications.” Over 350 IBM partners have developed cognitive powered apps out in the market today ranging from industries like healthcare to financial services to retail.

A Fluid Shopping Experience

A company called Fluid has used Watson to create the Fluid Expert Personal Shopper app designed to help consumers find the products they want more quickly at a retailer’s website. For questions about whether you want a wind-resistant coat, for example, the Fluid app helps you shop via a series of interactive panels based on whether the temperature will fall below freezing where you’re likely to wear it.

Steve Abrams, a Distinguished Engineer & the Director of Technology Solutions for the IBM Watson Group, says Fluid is being used both in-store and to enhance an ecommerce site. “You log in with your social media and it ingests what you’ve written and Watson has a personality insight, and Watson has a personality matrix covering 52 different dimensions of personality.” From all that data, he says, the app gleans such things as whether you’re more introverted or extroverted and then suggests products based on that analysis.

On the sport side, some pro teams are using Watson to enhance fan engagement. Log in at the stadium and Watson will engage you with relevant blogs and fantasy sports options you can play during the event. It will even provide comments on certain sports writers, noting, for example, if they’re considered biased.

Amazon Machine Learning

As for Amazon, analyst Charles King of Pund-IT says it’s doing a great job of making predictive analytics built on its Amazon Web Services (AWS) easy to use and accessible.

“Amazon offers a simplified platform for developers who want to start working with machine learning without a lot of stress or specialized tools or investment,” King, Principal Analyst at Pund-IT, told Datamation. “Typically machine learning has required developers to have a lot of specialized training, and required businesses who want to use it to invest in software tools and specialized algorithms and the hardware to support extremely large data sets.”

Amazon customer BuildFax tracks housing remodels and new commercial construction for insurance companies. BuildFax uses Amazon Machine Learning (ML) for predictive modeling and gives roof-age and job-cost estimations for insurers and builders. The company uses data sets from public sources and from customers to build models.

With Amazon ML, models that previously took BuildFax six months or longer to create are now complete in four weeks or less. The company also says Amazon ML creates opportunities for new data analytics services BuildFax can offer customers, such as text analysis to estimate job costs with 80 percent accuracy.

BuildFax CTO and founder Joe Emison says Amazon ML democratizes the process of building predictive models. “It’s easy and fast to use, and has machine-learning best practices encapsulated in the product, which lets us deliver results significantly faster than in the past.”

Jack Gold, Principal Analyst with J.Gold Associates, says IBM and Amazon are among those at the forefront of making AI more accessible for real world enterprise applications and it won’t be long before it’s more commonplace.

“What IBM is trying to establish with the Watson analytics engine is not just storing and acquiring data, but taking all that information and doing something meaningful with it as an AI service or Intelligence as a Service,” Gold tells Datamation. “Ultimately everyone is going to want to take advantage of these kinds of systems because companies have so much data they don’t know what to do with.”

No Big Iron Investment Needed

For budget-constrained enterprises, Gold says IBM’s approach is attractive because they don’t need to buy and maintain their own mainframe to leverage Watson’s technology.

“IBM understands that they are not a hardware company per se anymore, that’s not where the future lies with them. The real money is in services, and a lot of Watson has ties to professional services,” says Gold. (Of course, you don’t need to buy hardware for Amazon either, which has been a pay-as-you-go service from the start).

Ultimately, Gold says enterprises will inevitably use deep learning systems like Watson (or Amazon or systems from Google, Microsoft and others) because they have more data than they know what to do with. As these systems evolve, they will get better at extracting meaningful insights from all the data being collected.

The Deep Learning Road Ahead

Gold says we’re at a very early stage with these systems and IBM is smart to get developers involved, creating solutions that expand Watson’s appeal. “They’re providing tools on top of Watson, but building out the ecosystem is going to take a while,” he said.

IBM’s Abrams says there are a lot of hot areas in terms of developer interest. “Medical has been strong and growing. We launched the Watson Health Group a few months to capitalize on the interest as a new business unit,” he said.

Asked what is the best Watson app so far, Abrams defers. “It’s the one that hasn’t been written yet. The growth trajectory and partner interest we’re seeing is unbelievable. This time next year I’m sure we’ll see a new set of applications we didn’t anticipate.”

Big data and social technologies analyst Andreas Weigend says these new deep learning systems will have a lot of implications for retailers and other industries going forward.

“Think about facial recognition and emotion recognition. If you are a manager in a retail store, you would be able to tell know which staff members are doing well because you have a system that can read the facial expression of clients,” says Weigend. “I think we’ll see more of how work gets done and systems that decompose work. The notion of education and work hasn’t changed that much in the last thousand years, but stay tuned, change is coming.”

Photo courtesy of Shutterstock.