Folks in Paris have figured out how to more than double quantum memory storage.
Researchers at Laboratoire Kastler Brossel (LKB) in Paris have broken through a key barrier in quantum memory performance. Their work has enabled the first secure storage and retrieval of quantum bits.
Physicists at LKB have more than doubled the efficiency of optical qubit storage—from 30 percent to close to 70 percent—making secure storage and retrieval possible. Quantum memory is essential for future quantum networks. The ability to synchronize quantum bits has applications in long-distance quantum communication protocols or computing algorithms. With efficiency at well over 50 percent, quantum storage now enables protocol security.
In the January 25th online issue of Nature Communications, Prof. Julien Laurat and his team at the LKB, part of Sorbonne University and the CNRS, reported that they have stored optical qubits with a record efficiency of 70 percent, while preserving a fidelity to the input qubit beyond 99 percent.
“We selected a number of key elements and were able for the first time to combine them in a single setup. This work was key to achieving the highest efficiency to date for the storage and readout of an optical qubit,” says Pierre Vernaz-Gris, a former doctoral student who performed the experiment and one of the two lead authors of the paper.
The experiment involves the conversion of a photonic qubit into an atomic excitation of laser-cooled cesium atoms. With the protocol of electromagnetically-induced transparency, a control laser beam makes the medium transparent and slows down the impinging signal light carrying the information. When the signal is contained within the ensemble and the control beam is turned off, the information is converted into a collective excitation of the atoms, which is stored until the control beams is turned on again.
Hello, quantum money and communications!
I found this 10 minute video from August 2017 about farming tech . Lots of advances, and the last couple of minutes show autonomous farming bots.
Finally, an end to humans facing the tedium of jogging is in sight.
In case the terror of watching that isn’t enough, know that SpotMini can now autonomously navigate stairs , in addition to being able to open doors .
I’d like to see a video of that thing lying on its back, waving its legs in the air.
Check out this robot/waldo, straight out of Hinokio .
When I saw the case I thought “Baymax!” but it was the controllers.
Very, very cool thing there. I can see it changing doctor’s lives immensely. Not to mention tourism being completely re-made.
This just popped up on my AP feed: Maker of fearsome animal robots slowly emerges from stealth !
It’s a fairly brief article about Boston Dynamics; I highly recommend it if you’re interested in these things (and why else would you be in this thread, eh?).
Snowboarder_Bo:
Behold! The AI that can make fake videos!
I know the resolution is pretty bad (like a dashcam from 10 years ago), but the images are amazing. If this was offered as evidence in a court case, I doubt the jurors would be able to tell which was real and which was fake. Something like this could make for some troubling developments in the future.
Snowboarder_Bo:
There was a good article at Motherboard earlier this week*: AI-Assisted Fake Porn Is Here and We’re All Fucked Add in Adobe’s new software that let’s you make people say things via an algorithm that takes less than 20 minutes to mimic anyone.
Now add in the software that can make anywhere be anytime .
:eek:
I think our definitions of privacy and trust are about to change a lot .
*ETA: I see on re-read that that article links to the Motherboard article; nice!
This was posted just 3 weeks ago : are you sure you saw that person say or do that? Brief article here which also has the video embedded .
This is nearly flawless. At times it IS flawless.
I saw this on the AP today and thought it fit for this thread; it’s pretty ddamned amazing, IMO: IBM computer proves formidable against 2 human debaters . Debated subjects in real time, by voice.
The computer made its case for government-subsidized space research by pulling in evidence from its huge internal repository of newspapers, journals and other sources. After delivering opening arguments, the computer listened to a professional human debater’s counter-argument and spent four minutes rebutting it.
The company unveiled its Project Debater in San Francisco on Monday. IBM selected possible topics based on whether they were debatable, but neither the computer nor the human debaters knew the topic in advance. Nonetheless, the computer championed the topic fiercely with just a few awkward gaps in reasoning.
IBM doesn’t try to declare a winner of the debates, but Noa Ovadia, one of the human debaters, said the computer was a formidable opponent even if it made a few too many blanket statements about space exploration being the pinnacle of human achievement.
Ovadia, a national debate champion in Israel, said she was impressed by its fluency in language and ability to construct sentences. She said the computer was able to “get to the bottom line of my arguments” and respond to them.
Among several outside experts IBM invited to attend Project Debater’s debut was Chris Reed, who directs the Centre for Argument Technology at the University of Dundee in Scotland. Reed said he was impressed by its grasp of “procatalepsis” — a rhetorical technique that involves anticipating an opponent’s argument and preemptively rebutting it.
Pretty nifty, eh.
AIs now exist that work together to achieve a common goal . This is huge, IMO.
The manner in which these bots cooperate with mates cannot be taken lightly. Will Knight in MIT Technology Review: “This is an important and novel direction for AI, since algorithms typically operate independently. Approaches that help algorithms cooperate with each other could prove important for commercial uses of the technology. AI algorithms could, for instance, team up to outmaneuver opponents in online trading or ad bidding. Collaborative algorithms might also cooperate with humans.”
They tested the AI5 against an amateur human team and won, but they plan to play against a top human team on 28 July. I have no doubt they’ll destroy all humans, just like the single AI destroyed the world’s best player last year.
They decided to test the skills against amateur human teams. She said they were surprised: They have won their first games. While they were pleased, they still went to an expert to take a look at their work.
William Lee, better known as Blitz in the player community, is that person, regarded as quite good at the game. He said it took him eight years to learn some of the strategies that the bot was intuitively doing. He said the team fight bot aspect was excellent. Didn’t mess up. (Simonite mentioned that “Each day, OpenAI Five played the equivalent of 180 years of Dota 2.” He reflected on that. “No human has 180 years to learn a videogame.”)
So what happens after they beat a top human team?
The world championships will take place in August.
Dennison: “While the best players in the world are getting ready to compete, we’ re also working on the next version of our bot.”
I think the AI have teamed up to change your link to Carpool Karaoke
Well that’s a pretty good bit of technology, too, so at least it’s not a complete waste of code.
Here’s the proper story: Dota 2 challenging bots turn hard to beat after being taught cooperative mode
I admit I am unclear on how that is an example of multiple AIs working together.
DataGrid created a GAN (generative adversarial network), which consists of two separate neural networks. The first network fed facial images of existing, real-world idol singers into the second network, which then analyzed them to determine the special characteristics that constitute an “idol face.”
You may wish to call the two neural networks something other than AIs. I’m not going for that level of hair-splitting.
That is not even remotely the same concept of “working together for a common goal” as what I mentioned; it is merely an example of reinforcement learning. The two bots you’re talking about don’t make decisions to enable the other to better perform a task or to re-divide the tasks they are performing to optimize an outcome, they are each simply doing the job they were designed for. They aren’t “working together” any more than two cogs are “working together”.
Seriously: robot stuntmen !
Did you see that superhero pose in the last shot? I don’t see why you wouldn’t use these and then just plaster people’s faces on them in post! You could do some outrageous stunt shots!