Select Page

This week in AI #15: Which side are you on?

by | Tech

Did you miss your holy AI Hajj pilgrimage this year? We’ve been to the frontline and came back with lots and lots of ideas. For a better or worse world, you’ll get to choose your team!


1 — Cool Kid

No worries, there’s still a long way to go before we build kids such as David from A.I. by Spielberg.

The new generation is here. Forget X, Y, Z; the AI gen is coming for you. Don’t believe me? Look at this 22-year-old from Shanghai that just won first place in a rear vehicle detector challenge! 22 is too old you’re telling me? What about publishing a research paper with OpenAI at the ripe age of 17? Yes, that’s what I thought. Alright, alright, I might have fallen into a bit of media sensationalism. But even though those two are exceptional individuals it’s an encouraging sign to see the AI community opening-up. If only for two reasons. First, the more people develop this particular skill set, the more we’ll be able to embed AI in our everyday life and change our society for the better. And second, the broader the community becomes, the more diverse it gets!


2 — O Sole Mio!

A picture of ICCV 2017 snatched by one of our three emissaries in Venice while we were hardworking.

While we’re speaking of research papers, this week was very special for all the AI fans out there; the International Conference on Computer Vision took place in Venice. With NIPS and CVPR it consists the Holy Trinity of AI. The three religious events a zealot cannot miss. All prominent researchers are presenting their latest breakthroughs, and all tech giants are there too. It’s the place to be. This year didn’t disappoint: from one-pixel attack to fool neural networks, to unified convolutional networks for detection and tracking, or even visual reasoning, no matter your taste, there’s something for you!


3 — You look familiar…

Look at the pictures above. Do you know who they are? Hard to put a name on those two? That’s okay. A machine generated those faces. In itself, this isn’t newsworthy, GANs have been around for some time now, and we’re used to what they’re capable of. Although, that’s only at first glance. If you look closely, something is a bit different this time. You see, GANs have been notoriously known for being great at low resolution but not scaling up. What this means is that if you wanted to generate real-looking high-res pictures, GANs weren’t cutting it, which was a notable drawback. Well, no more! The research team at Nvidia just published a new approach to training them from the coarser to more delicate details plus tons of additional tips and tricks! It might prove to be a game changer; one day maybe, generating samples will be one of the standard data augmentation processes, such as horizontally flipping an image!


4 — Careful what you wish for

Generating new faces may come in handy sooner than we think! There’re a lot of fears concerning AI especially when it comes to identifying individuals and tracking them: through their social media interactions, their commutes, their purchases, and even their movements! What a dystopian world that would be! Fortunately, we’re not there yet at all! Or are we? No, really, are we? Well… in a sense, yes. All those examples I quoted earlier? They didn’t come out of my imagination. A real artificial intelligence system called City Brain is currently gathering this data in Hangzhou. And it’s using it with high efficiency, reducing crime, car crashes and rush hour traffic. The giant behind this tech is Alibaba, and it made AI one of its top priorities, going as far as investing $15 billion in its DAMO program!


5 — Team Neanderthal vs. Team Singularity

Listening to the Zucks and Musks of this world, it’s easy to think that there is only one truth: either we embrace AI, and it will enable us to transition to a type III civilization, or we realize the danger it represents and go back to farming and hunting. As always, the world is not black or white. Resorting to such a Manichean view only leads to both sides entrenching themselves and the debate becoming sterile. That’s why it’s always a good idea to get a glimpse of the driving forces behind the two opposite factions. So this week, if you want to dive deeper on the positive side, we recommend Life 3.0: Being Human in the Age of Artificial Intelligence by Max Tegmark.


That’s all for this week. Did you enjoy reading it? Because we sure do love writing it! If you appreciated it, make sure to give us a clap and subscribe to the newsletter below 👇👇👇 ! Any feedback? Let us know in the comments section!

https://upscri.be/edd62c/

Share This