Subscribe

May 30 2017

Can't Quite Kick That Human Habit

Discovering the world of artificial intelligence with Axiom Zen.
Can't Quite Kick That Human Habit
It turns out spiders are good for more than just eating flies and terrifying the unwary—they're also providing fascinating insights into the way non-human creatures might think. It turns out, not everything is as brain-focused as we are (if these researchers are right, even we might not be as brain-focused as we think we are).

Plus, AlphaGo is quitting while it's ahead, Facebook can't quite kick its human habit, and Blippar is smarter than you.
Article of the Week
The Thoughts of a Spiderweb
Mapping neural networks based on insect behaviour patterns is a branch of machine learning called swarm intelligence. It's pretty common, and often effective, but we're now learning that insects may use more than just their brains to think. This brings up the 'body problem,' the idea that researchers put aside the brain-body connection as too complicated, and never went back to it.

"Many animals interact with the world in certain complicated ways that don’t rely on their brains. In some cases, they don’t even use neurons."
The Bleeding Edge
This startup thinks it can predict whether or not your current relationship will last - just answer six questions and they'll give you the percentage chance that you'll make it through. The algorithm seems painfully simplistic - one of the questions, for instance, is how long you've been together (the longer a couple is together, the less likely they are to break up). But they plan to use the data to help build out their machine learning capabilities.
AlphaGo Retires From Competitive Go After Defeating World Number One 3-0
Once is a fluke, but twice is apparently enough evidence for DeepMind to retire AlphaGo, calling it an overwhelming success. It's up to others to decide how to turn this success to the human good—DeepMind is on to the next exciting challenge, though no word yet on what that will be.

“More than a competitor, AlphaGo has been a tool to inspire Go players to try new strategies and uncover new ideas in this 3,000 year-old game.”
Facebook has faced a lot of scrutiny over their inability to filter problematic content—from livestreamed incidents of violence to fake news. They're hoping to use AI to help flag content, but right now those computers are just doing their best to make life easier for human moderators, which is still where the puck stops.

"Importantly, in the end, the determination as to whether or not to actually make the report is left up to people. The AI just can't cut it alone, but it's helping."
Blippar’s Machine Learning Tech Can Identify Cars Better Than You Can
Machine vision is a key step for augmented reality—the technology needs to be able to understand the world around it, and that means some pretty sophisticated artificial intelligence. Cars are a great dataset to train on, and Blippar is doing exactly that. Their AR can now differentiate between makes and models of cars better than most humans.

"In other words, Blippar’s AI can identify the make, model and year of any U.S. car made in 2000 or after, as long as the car is traveling slower than 15mph."
In the past, Uber fares were generated based on mileage, time, and geographic demand. The new "route-based pricing" system, on the other hand, uses machine-learning to charge whatever they think they can get away with.

"Should a passenger book a ride from one wealthy area to another, for example, they may be asked to pay more than someone heading to a less affluent area."
“By far the greatest danger of artificial intelligence is that people conclude too early that they understand it.” ― Eliezer Yudkowsky
AI@AZ · #350-980 Howe Street · Vancouver V6Z 0C8 · Canada
Unsubscribe
Archive