There have been some exciting developments in the world of Machine Learning and Artificial Intelligence! Let’s check out some of them:

Robot pizza delivery

What better way to start off this blog than some cheesy news? That’s right, Dominos and Starship plan to have robots deliver pizza in Europe! Domino’s once brought this up before, with their “Domi-No-Driver” campaign in 2015. 

While it might have been a joke then, it’s clear that they’ve been pursuing it ever since, because it’s all set to go live in Hamburg, Germany in a few months. Dominos has signed off on the dotted line with Starship Technology to use their “personal delivery devices”, and we’re very excited to see how it goes. Meanwhile, in New Zealand, Domino’s has begun delivery by drones in a partnership with Flirtey.

Google’s AI invents its own secret language

Rub your eyes, because you read that right! Google has an AI program called ‘Google Machine Neural Translation’ (GMNT), which was created to assist the Google Translate application. The plan was to have GNMT learn from Google Translate requests to create more natural sounding translations. 

Now here’s where it gets weird. If GNMT was programmed to translate from Hindi to English and from Kannada to English, it can translate from Hindi to Kannada (something it wasn’t programmed to do) by using a self-created secret inter-language or ‘interlingua’. The scary part? Google’s AI researchers have not decoded the Interlingua yet.

‘Tree – on – a chip’ passively pumps water

Engineers at Massachusetts Institute of Technology (MIT) have created a chip that mimics natural water mechanisms in plants. Trees pull up water from their roots to the tips of leaves, and pumps sugar generated from the leaves back to their roots, and this nutrient shuttle is constant. 

Now MIT’s impressive chip is trying to emulate this efficient natural process, and has managed to get both sugar and water pumping through the chip for several days without using an external motor or pump! This has huge implications for the motor industry worldwide, especially when it comes to robots which rely on large expensive motors to power their movements.

‘Machine Folk’ music successfully composed by AI system

Researchers from London have trained an AI system using more than 23,000 ABC transcriptions of folk music. These transcriptions have been crowd-sourced from the online resource thesession.org, to create a system that can generate folk music by itself. Their AI is trained such that when it is given one ABC symbol, it can predict the next symbol, thereby generating patterns and sequences of notes and pitches that make music ‘music’. While the music received mixed reviews from listeners, trained musicians said 1 in 5 of the tunes were good, and with the system having already generated 10,000 tunes, it won’t be long before we hear some on the radio. 

So what are some of the other exciting developments that you’ve come across in Machine Learning and Artificial Intelligence? Tell us via comments.