UPDATED 19:20 EST / JUNE 23 2017

BIG DATA

Hilary Mason sees progress in machine learning, but it’s not ready for prime time

IBM Corp. is making big bets on machine learning, with recent announcements about moving its cognitive computing platform Watson to the mainframe and combining its Data Science Experience  with Hortonworks Inc.’s big data management platform. However, as one prominent computer researcher told a gathering of developers and information technology professionals, while advances are clearly being made, there is still a long way to go before machine learning technology can be confidently implemented across platforms on a global basis.

“We’re starting today to see real value coming out of machine learning. But it’s very difficult,” said Hilary Mason, famed data scientist and founder and CEO of Fast Forward Labs Inc.

Mason spoke at the IBM Fast Track Your Data event in Munich, Germany, this week and offered her views on the current state of the data science world. Her company produces reports each quarter on new technology breakthroughs and builds data science prototypes.

“We aim to be about six months to two years ahead of what’s in the market,” Mason said.

Language processing making strides

As evidence of emerging value, Mason cited a couple of examples that she has worked on recently. One involved research in natural language generation technology where small amounts of structured data are ingested to produce computerized articles in narrative.

“If you hear about this at all it’s because reporters are freaked out because they think it’s going replace their jobs,” said Mason, who doubts that would actually happen. “It is about making data clear to people through language.”

Her prototype has been used to take data about a New York City apartment and write an advertisement listing it for rental. Mason also has a client who is using it to scan information about major movie and TV stars and generate articles in the realm of celebrity gossip.

There are also signs that machine learning is beginning to scale. Mason described how the prototype was applied to a corpus of 10,000 Amazon product reviews and then summarized down to only 10 primary points of view.

“You start with something you can do anyway and then find an ability to scale that to something that you could not do as a human being before,” Mason said.

Recognizing photos and video still difficult

Although neural networks are advancing, still photo recognition remains a thorny problem. Mason described results from an Instagram project where machine learning fell a bit short. A computer program combined subways with prisons and crabs with French fries.

The Fast Forward Labs founder has encountered similar issues when neural networks are applied to video footage as well. Being able to associate movement with context was very difficult, and she pointed out the challenge of monitoring “a video of someone working out and a video of ISIS” and having a machine understand the differences.

“That’s actually very difficult to do in a purely automated way,” Mason said.

Designing algorithms to solve these and other problems has been a complicated challenge, especially when it comes to the field of healthcare. Mason referenced a recent example where an algorithm recommended that U.S. patients sent to the emergency room after having an attack should be sent home because the data presented evidence that showed that asthma suffers got much better care than others.

“Of course, in healthcare it is critical, but in classifying pictures of cats and dogs, no one really cares,” Mason concluded.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s independent editorial coverage of the IBM Fast Track Your Data event. (* Disclosure: TheCUBE is a paid media partner for IBM Fast Track Your Data. Neither IBM nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE

A message from John Furrier, co-founder of SiliconANGLE:

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.  

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU