Machines are getting better at understanding what people say, but they’re not perfect.
There’s still quite a bit of progress to go before they fully ‘get’ the complexities of our language. In a TechCrunch article from this week, however, that might change sooner than we think.
A new system dubbed “ELMo” is adding context to words that had previously stumped machines. To illustrate the problem ELMo solves, consider Devin Coldewey‘s example:
“To illustrate the problem, think of the word “queen.” When you and I are talking and I say that word, you know from context whether I’m talking about Queen Elizabeth, or the chess piece, or the matriarch of a hive, or RuPaul’s Drag Race,” said Coldeway in the TechCrunch article. “Which meaning it is can usually be reliably determined by the phrasing — “God save the queen!” versus “I saved my queen!” — and of course all this informs the topic, the structure of the sentence, whether you’re expected to respond, and so on.”
You and I can fill in the context in the above example, but computer systems can’t. They don’t quite have that level of flexibility.
ELMo, or Embeddings from Language Models, however, helps tackle this issue by using “its training data (a huge collection of text) to determine whether a word has multiple meanings and how those different meanings are signaled in language.”
Matthew Peters, lead author on the paper that details ELMo, said the goal of the solution is to reduce human annotation.
“We were looking for a method that would significantly reduce the need for human annotation,” explained Peters. “The goal was to learn as much as we can from unlabeled data.”
What to Expect for the Factory of the Future
This week, IndustryWeek posed the question, “What makes a factory of the future”? And how do we know it’s just around the corner?
After all, people have been searching for the Factory of the Future since the 1880s. It’s a never-ending job.
“Manufacturing plants and equivalent facilities like oil refineries and power stations stand as remarkable achievements in engineering and process innovation, but they are also complex, multifaceted and can always be improved,” said Michael Kanellos in IndustryWeek.
Kanellos says there are at least three key expectations for the Factory of the Future.
“Energy can take up 30% of the cost of producing metals or running a water treatment plant. Data centers spend more on energy than anything else except employees. Renewables have become a high priority for many data centers because (1) it helps the company reputation and (2) it can turn what can be an unpredictable, variable cost for 30 years into a smooth, capitalized asset.”
Change in Ownership
“Rather than sell the equipment they manufacture, equipment makers will install systems, maintain and control them and sell access to them as a service. Equipment as a service creates some legal headaches but financially it makes quite a bit of sense for everyone.”
“As more systems get monitored and owned remotely, many employees (or rather the employees of your equipment providers) won’t need parking spaces. They will be working in control rooms off-premise.”
How GE Aviation Unlocked Revenue with Asset Performance Management
As part of GE Aviation’s digital transformation journey at their Muskegon facilities, the team worked to shift preventive maintenance from schedule-based to condition or machine use-based to ensure best application of key resources. This video explains how they developed a framework for continuous improvement by partnering with GE Digital to implement Asset Performance Management (APM) on its engine production lines.
For years, Transform, the GrayMatter conference in Put-in- Bay, Ohio, has been the place to meet for industrial organizations undergoing digital transformation. This year we’re expanding the event to showcase advanced digital industrial applications with a special track on Asset Performance Management (APM).
You’ll find sessions on:
- Finding hidden capacity at chemical companies like Eastman
- Reducing maintenance costs at a paper company by over a million dollars a year
- Maximizing resource planning by shifting to preventative maintenance from schedule-based to stop bad things before they happen
Transform 2018 is GrayMatter’s annual conference held in Put-in-Bay, Ohio. The event is designed for professionals in all verticals who are passionate about operational technology and transforming into digital, industrial operations.Attend Transform 2018