WTiN - Artificial Intelligence & Machine Learning

Artificial Intelligence (AI), Machine Learning (ML), the Textile Industry, surely not? Surely YES!

Here at WTiN we’ve been following this subject for some time; in fact, we’ve been using elements in our products for several years now, but we’ve got some exciting plans for future features, services and projects. So how are we are leveraging AI, ML and associated technologies and what are those future plans?

What?

What are AI and ML and the difference between them? Coined in 1956 by John McCarthy, AI involves machines that can perform tasks that are characteristic of human intelligence. These include things like planning, understanding language, recognising objects and sounds, learning, and problem solving. There is (and will always be) a debate around whether AI is truly possible. ML on the other hand is simply a mechanism for achieving AI. ML is:

“The ability to learn without being explicitly programmed.”

Usually, this involves ‘training’ algorithms using large quantities of historical and/or known data to enable prediction (with a certain level of probability) of a future outcome based on changing inputs. This sounds like a classic predictive model, but the difference with ML is that a system can continually update the model and ‘learn’ on the job. The two terms are not interchangeable and should not be used as such, so I’ll stop doing it!

Why?

Something we’re keen to stress is that we’re not looking to use new technologies just because we can. AI tools mixed with some of our own proprietary ML are and will add value to WTiN’s products:

  • Understanding – What does content mean and what is it about? By using AI to extract content metadata (see the ‘how’ section below) we can analyse, link and extract insight from content like never before.
  • Aggregation – We’ve been aggregating content using AI tools for a long time, but technology is now allowing us to use this content in many new, truly valuable ways.
  • Sentiment – Understand what the World is thinking about the textile manufacturing industry, a sector of it, a company in it or even an individual key opinion leader! Is it positive or negative?
  • Trend – How are things moving and what are the trends? Knowing what’s happening now is useful, knowing how things have changed and combining this with WTiN’s insight and ML capabilities makes for a very powerful predictive model.
  • Discoverability – Overarching the above is the opportunities the technology offers to our customers to discover more information of new types in the most efficient way. Finding the micro detail, not just the macro!

How?

So how are we doing this? We’re using several tools and processes. Here’s just a few.

IBM Watson
One of the major tools in our arsenal, which is why you’ll see logos on wtin.com and on WTiN email communication. We use IBM’s ‘Watson’ natural language understanding service to process all of our content. Below is a video from IBM about the service that you might find interesting. It’s general and not specific to our industry, but it gives a flavour of how we use the service to extract metadata from our content. We then use this data to help build-out some of the services we’ve already talked about.

Microsoft Machine Learning
The big ‘M’ provide a great platform for building out machine learning, automating models and removing some of the monotony associated with building predictive ML models. This official line sounds pretty impressive:

“Build artificial intelligence (AI) applications that sense, process,
and act on information—augmenting human capabilities, increasing speed
and efficiency, and helping you achieve more.”

We use MML to help learn from our content and sectorise it. When we aggregate content from all over the Internet MML helps us automatically determine which of many WTiN domain specific sectors content belongs in, and it’s learning all the time!

You can find out more about the platform here - https://azure.microsoft.com/en-gb/overview/machine-learning/

Custom R coding
We also utilise some custom developed R code to complete data analysis across several areas of our content.

In truth we use lots of other tools in the process and we’re developing all the time. Keep listening for more news and product developments and if you’d like to learn more then feel free to get in touch!

Steve

Steve Ayres
Head of Media and Technology
sayres@wtin.com