salto jumping robot-1481045910872
UC Berkeley’s Salto Is the Most Agile Jumping Robot Ever
February 9, 2017
astro head shot 1240 v2-1480719986090
Astro Teller, Captain of Moonshots at X, on the Future of AI, Robots, and Coffeemakers
February 9, 2017

Video Friday: Kengoro the Sweaty Robot, Camera Drone on a Leash, and the Next Frontier in AI

kengoro sweat robot-1480701831510

Video Friday: Kengoro the Sweaty Robot, Camera Drone on a Leash, and the Next Frontier in AI By Evan Ackerman and Erico Guizzo Posted 2 Dec 2016 | 18:01 GMT

Image: Kazumichi Moriyama via YouTube
Kengoro is looking pretty beefy.

Video Friday is your weekly selection of awesome robotics videos, collected by your Automaton bloggers. We’ll also be posting a weekly calendar of upcoming robotics events for the next two months; here’s what we have so far (send us your events!):

Humans, Machines, and the Future of Work – December 05, 2016 – Houston, Texas, USA
RiTA 2016 – December 11-14, 2016 – Beijing, China
SIMPAR 2016 – December 13-16, 2016 – San Francisco, Calif., USA
WAFR 2016 – December 18-20, 2016 – San Francisco, Calif., USA

Let us know if you have suggestions for next week, and enjoy today’s videos.

Back in June, IHMC posted a video of ATLAS walking on the edges of rotated cinder block all placed in a line. This latest video shows the robot walking on the edges of cinder blocks that have also been rotated in the horizontal plane, giving the robot’s feet less surface area to work with:

There’s one more dimension to add before ATLAS will be walking on randomly positioned cinder blocks, and we’re expecting to see the real robot handling point contacts soon as well. And then doing a crane kick from them.

[ IHMC ]

You don’t have a Cozmo yet, but when you get one, you can immediately upgrade it to be able to recognize dogs and cats. Not different dogs and cats, we’re pretty sure, but still:

And obligatory:

Is this how most robots feel?

[ Anki Cozmo ]

Turtlebot 3 is getting some cool stuff, like a bajillion ways to teleoperate it:

Can I buy one yet? Please?

[ ROBOTIS ]

Kengoro is looking pretty beefy, even if it does need a giant fan to (presumably) keep its arm cool. Better get that sweatin’ on!

And from the same iiiEx show in Japan, there’s this, which I think might be a robotic tardigrade?

[ iiiEx ] via [ Kazumichi Moriyama ]

The Fotokite Phi, which we featured on our Robot Gift Guide this year, is still technically a drone, but we really like the fact that it’s very safe and very easy to control:

[ Fotokite ]

I wasn’t aware that Ford was using ABB’s Yumi collaborative robots in vehicle production, but they’re certainly teaching one to be a DJ:

Something about a new car, too.

[ ABB ] via [ Ford ]

This video with a Roomba and a kitty is a little more action-packed than most videos of Roombas and kitties.

Roombas I’ve used at home (a 980 and an 880) have both tossed themselves down my stairs, which are worse than these. In one case, it wasn’t my fault, and in the other case, it was. But both robots are somehow, miraculously, totally fine. iRobot build quality FTW.

[ iRobot ] via [ Laughing Squid ]

Hinamitetu has upgraded his gymnastics robot with a “voltage stabilizer,” and suddenly it can nail all of its tricks:

[ YouTube ]

I saw this on Facebook and thought it was cool:

Iron Man, right? It’s 3D printed out of plastic, with an integrated battery, controller, and three motors.

[ 3DPPMP ]

CERN’s TIM (Train Inspection Monorail) patrols the 27-km tunnel of the Large Hadron Collider making sure nothing is amiss:

Those little monorail doors are just adoorable.

[ CERN ]

Most of us have no idea how to program a robot to pick up a specific object, but like all problems, this one is solvable with lasers.

I assume there is a really sad Baxter somewhere with only one arm.

[ UMass Lowell ]

MekaMon’s battle quadrupeds can now entertain you in augmented reality, if you have no friends (or no friends with other MekaMons):

$250 for one mech (mek?), or $400 for two.

[ Reach Robotics ]

A DARPA/ONR/Northrop Grumman team is developing Tern, a revolutionary medium-altitude, long endurance, highly autonomous unmanned air system. The new tail-sitter UAS has the potential to dramatically improve the quality, capability, and reach of future Navy ISR and strike, and Marine expeditionary operations around the world.

[ DARPA ]

This is a recently posted video of Ray Kurzweil interviewing the late Marvin Minsky, who died early this year at age 88. In the clip below, they talk about human-level AI. For the full interview (it’s not clear when it took place), hit the link below.

[ Full Interview ] via [ AI ]

I could listen to Rodney Brooks talk about anything forever. It’s either his accent, the fact that he knows tons about robotics, or some combination of both.

[ Rethink Robotics ]

Why toddlers are smarter than computers: Advancements in artificial intelligence are changing how we analyze and process information. But these advances fall short when compared with the ingenuity and creativity of an average three-year old. In his talk, psychologist and neuroscientist Gary Marcus compares advancements in AI to the resilience of the human brain on tasks such as common sense and real-time evaluations.

Thanks for the shout-out about our DRC robot falls compilation Gary!

[ TEDxCERN ]

And one more AI video! The CMU RI Seminar features Yann LeCun, director of AI research at Facebook and professor of computer science at NYU (check out our in-depth interview with LeCun from last year).

The Next Frontier in AI: Unsupervised Learning

The rapid progress of AI in the last few years are largely the result of advances in deep learning and neural nets, combined with the availability of large datasets and fast GPUs. We now have systems that can recognize images with an accuracy that rivals that of humans. This will lead to revolutions in several domains such as autonomous transportation and medical image understanding. But all of these systems currently use supervised learning in which the machine is trained with inputs labeled by humans. The challenge of the next several years is to let machines learn from raw, unlabeled data, such as video or text. This is known as unsupervised learning. AI systems today do not possess "common sense", which humans and animals acquire by observing the world, acting in it, and understanding the physical constraints of it. Some of us see unsupervised learning as the key towards machines with common sense. Approaches to unsupervised learning will be reviewed. This presentation assumes some familiarity with the basic concepts of deep learning.

[ CMU RI Seminar ]

The entirety of the 2016 Bay Area Robotics Symposium is now online in the form of one massive 9-hour-long recorded livestream. Our recommendation: check out the schedule and the link below, do some simple math, and just skip directly to the talks that you want to see.

[ BARS 2016 ]

Learn More

robot videos video friday