Good morning, my little cyborgs!
Welcome to the latest Sunday AI Review. Hope you’ve all had a fantastic week. I’ve certainly had a busy one, so am really enjoying a nice, relaxing weekend, just me and Indie…
But hey, it’s that time of the week again where I come bearing all the news of this week in artificial intelligence. Before we get going with the news, make sure you check out this list of the best artificial intelligence events in 2018. Jonny and I spent ages pulling this together to offer you a great resource to help you plan your events calendar for the year, so do please enjoy.
Oh! And if you’re an events organiser, please do drop me an email and let me know if you’re not on the list and I’ll get you added in straight away!
The most socially-shared article this week is a piece by Microsoft’s Brad Smith, President and Chief Legal Officer, and Harry Shum, Executive Vice President of Microsoft AI and Research Group. The article introduces the authors’ new book (same title as article), summarising the issues that the book seeks to address. What’s interesting is the emphasis on ethical considerations around AI, which – for me – always feels like a bit of an Easter Egg insight when it comes from one of the overlord corporations.
So, Microsoft and Alibaba’s AIs have beaten humans at reading comprehension on the Stanford University test. Of the two, Alibaba came out on top, marking a ‘great start for 2018 [for AI]’ according to Pranav Rajpurkar (a Stanford researcher behind the reading comprehension test).
Is AI development more profound than that of electricity or the discovery of fire? Pichai certainly thinks so, and it’s pretty hard to disagree. The article is basically a teaser for an interview between Pichai, YouTube CEO Susan Wojcicki, and Recode’s Kara Swisher/MSNBC’s Ari Melber which is set to air this coming Friday on MSNBC. UK viewers will have to sniff it out online somewhere.
Elon Musk and OpenAI Want to Create an Artificial Intelligence that Won’t Spell Doom for Humanity (Interesting Engineering)
We know how ardent Musk is about avoiding the robo-pocalypse, so this article is, frankly, much of the same news. It’s a good mission to be on, the whole ‘avoiding dangerous artificial superintelligence’ thing, so it’s important to note that, though he’s the one shouting loudest, ol’ Elon isn’t alone in his beliefs, nor his quest for safer AI.
Human art will never die, so let’s not get into that redundant discussion. It’s interesting to see that an AI is now able to use a kind of ‘imagination’ to draw things based on text descriptions. Sounds like the machine learning model being used is a GAN (generative adversarial network) which is quite a fascinating technique, one that pits two AIs against each other to advance the network’s learning process. I like what Microsoft’s got, but it’s still not as impressive as Google DeepDream’s imaginary visualisations, in my opinion, though I do appreciate the different functions of the two – Microsoft’s having more practical applications.
Pick of the Week
And one for luck… it’s not AI, but it is really interesting so here you go:
That’s all for this week, folks!
Have a great Sunday, and I’ll catch you again next Sunday for more from the weird and wonderful world of artificial intelligence news.