Machine Learning goes to the masquerade


By on October 17th, 2019 in Events, Main Blog
Machine Learning goes to the masquerade

Machine Learning (ML) has been around for much longer and is much more common than you think. What many people aren’t aware of is that it’s not reserved just for technical specialists or data savvy engineers - it’s there to help you and to make your life easier.

Chances are, you will have likely used it or at least benefitted from it at some point in your day, you may well even be using it right now. So, for this year’s Wirehive100 awards, we’re showcasing the impact that Machine Learning can have on your journey to the awards.

First thing first, you’ve got to look the part, so in comes the AI suit!

Suiting up with The Drop – AI & ML for suit fitting

Now, we’re not talking ‘The Tuxedo’-esque AI suit or ‘Inspector Gadget go-go arms’, we’re talking about machine learning & predictive analytics to make a custom fit suit.

The Drop, an online, high-quality tailor, uses a combination of customer photos and self-measurements to create their suits. They wanted to streamline the customer experience and therefore developed a bespoke solution to detect and flag anomalies in user measurements, generating and validating sizes through customer photographs and requesting additional measurements from the customer.

The Drop are already seeing the benefits of their solution by enabling their workforce to focus of creating custom garments faster.

What about a custom fit mask I hear you cry?

What’s a masquerade without a mask? Here come the GANs

You may have heard about the AI painter or AI art? But what is the technology behind it? That would be a GAN, or a generative adversarial network, a recent innovation in machine learning that we could use to design a masquerade mask.

It’s not that much different from other forms of ML in the sense that feed the model with data and it spits out recommendations. We just feed the model with plenty of images and get it to design our own. We could use it to generate a host of content, even this article (or so I keep being promised). We'll cover GANs in more detail another time.

Why not try making your own with the drawing bot from Azure: http://drawingbot.azurewebsites.net

Or, if you’re in a rush and haven’t got time to play around, we can utilise the Product Recommendation algorithms that span the retail industry, and which you’ll easily encounter using Google Shopping or Amazon to choose a product tailored to you.

Now that you’re dressed the part, we need to get there, and if you haven’t picked up the theme already, this is where ML comes in handy with transport.

Your ML Carriage Awaits

The use of machine learning in transportation has been a very useful development. Since the birth of SatNavs, applications have evolved to include more complex routing algorithms and collect more data to optimise journeys and empower consumers.

It’s all too easy to overlook the power of Google Maps but take a minute to think about all the information at the palm of your hand; the ease of finding and navigating to your destination, route suggestions and en-route optimisation, traffic alerts, alternative means of transportation, just in case we want to walk to Thorpe Park. A lot of that is made possible with machine learning.

Not impressed? What about the ML driven efficiency afforded to you by other travel applications that could save you a few quid?

Trainline uses machine learning to predict future tickets pricing. The company deploy this feature into its app to update customers with price snapshots and let customers know when a ticket price is likely to increase. This helps you save on advanced tickets. That doesn't sound scary at all.

How about a room at the inn?

So, we’ve arrived, and we’re dressed to the nines, mask and all, but now we’re faced with too many options of where to stay.

Thankfully, through personalisation, ranking and pricing optimisation, we’re able to find and select suitable accommodation tailored to our needs.

AirBnB and Booking.com both employ machine learning to help recommend and rank options for users. This can be done through simple classifiers or decision-tree based criteria, and these models and insights continue to increase in complexity as we peruse accommodation options and create a profile for the platform to consider.

DJ-ing in our DJs

We’re all set and now we can kick back and enjoy the music!

With the help of the machine learning employed by music apps like Spotify, we’ve got playlists for all parts of the event.

We’ve got a ‘focus to write the blog’ playlist, a ‘roadtrip/ en-route’ playlist, a ‘getting ready playlist’, a ‘are you getting the picture’ playlist, a ‘one more just in case’ playlist and a ‘feeling worse for wear’ playlist, all which can be formed on personal song choices, similarities in genre or bpm, or even collective selections. They also employ NLP (natural language processing) to scour articles and blogs to create playlists on potential trends - let's see if we can get them to create a Wirehive100 playlist.

Good luck to all the finalists!

+ There are no comments

Add yours