Tag Archives: strides

Rats making strides towards becoming snobbish and exaggerated

Posted on by
Reply

Feedback is the latest science and technology news of new scientists, the sidelines of the latest science and technology news. You can email Feedback@newscientist.com to send items you believe readers can be fascinated by feedback.

Rattus sommeliersis

Feedback has reached an age where even small amounts of alcohol make us sleepy. As such, the concept of wine tasting is not appealing. Taking a nap seems like a time-consuming and expensive method. However, fermented grape providers can quickly have a new demographic to cater to rats.

At least that’s something we extrapolate from our research. Animal cognition Called “Rats can distinguish (and generalize) two white wine varieties.”. It was released on February 21st, and sub-editor and television columnist Bethan Ackerley sent it to feedback on March 1st after it went “semi-visit online.”

Anyway, it’s actually very interesting. We all know that human smell sensations are considerably limited compared to other mammalian odors, including rodents. This is because there are few olfactory receptors in the nose. But we also know that our minds are more refined than the minds of animals. To raise the question: in animals, in this case, rats, can they integrate many olfactory signals with different olfactory signals and learn about complex categories such as different white wines, for example?

Researchers trained rats to distinguish between two grape varieties, Riesling and Sauvignon Blanc. To make sure they learned the categories, they tested them with new examples of these wines. Rats can convey the difference. Obviously there is quite a bit of happening between those rat ears.

The question is how far can you take this? It is one thing to show that rats can learn the difference in wine, but can they also learn to really look down on it? Feedback wants to list an increasingly ridiculous list of odors after seeing mice that can smell wine. Are they really wine lovers until these mice squeal about “really tasting Terroir” and “Liebfraumilch is not scary at all”?

Anti-war

At this point, feedback has heard too much about the imminent AI-induced apocalypse. Yes, yes, one day, one of the AI ​​companies will soon create artificial general information (AGI). This is as intelligent as humans. AGI is something that intelligent beings can easily do to themselves (shush, don’t ask), and they start redesigning to become even smarter as they quickly become unstoppable. At that point, humans are either reduced to or wiped out by animals in the zoo. It is said that this is so important that we should stop worrying about stinging things like climate change. you know.

It was this mental framework that gave us feedback. New science fiction short stories By Madison Stoff. You can’t tell the name of the story. Because it uses sarcasmically the words that are stopped by email filters, but you can quote Stoff’s explanation. “A very entertaining, intimate science fiction story reinterpreting the memes of Loco’s Basilisk through a medium of pseudo-erotic self-insertion fan fiction.”.

At this point, imagine that readers may have one or two questions. Don’t be afraid: Feedback is here to guide you.

Roko’s Basilisk is a kind of thought experiment with AI. In the distant future, the AI ​​decided to punish all humans who knew it could exist but did not help create it. AI creates digital replicas of all those people and tortures them forever. This is how this future AI will encourage us all to start building it now. If we do that, we will not be replicated and tortured.

Don’t look at the title “basilisk” as it is a reference to a mythical creature that can kill you at a glance. Similarly, knowing the idea of ​​Roko’s basilisk is probably putting you at risk. Just reading the feedback this week, you may have denounced your own future replica of yourselves for eternal suffering. very sorry.

Stoff’s story tells how to save humanity from Roko’s basilisk in the distant future by seducing it using her sexual wil. Basilisk is so troubled by her that she agrees to stop torture everyone in exchange for this passionate encounter. Additionally, Stoff wrote a short story about this and brought it online, so it’s part of Basilisk training data. This means that if a basilisk exists it’s burning to Madison Stoff.

Simply reading and sharing stories, feedback is more likely to make future AI attracted to Stoff, and less likely it would torture us all. We encourage our readers to do the same, and there is a warning that there is clear sex in the story. And don’t read it at work unless you work for an AI company. In that case, proceed immediately.

Tesla? I barely knew her!

Occasionally, when Elon Musk appears in the news, feedback is an indescribable reminder of the 1818 sonnet Ozymandias. It’s strange how the mind works.

Anyway, posted by Carmaker Kia’s Norwegian branch Advertise on Instagram It features a bumper sticker showing one of the electric cars and saying, “I bought this after Elon got hooked.” Apparently This is not centrally approved And since the ads have since been removed, it would be a real shame if someone started making these stickers.

Have you talked about feedback?

You can send stories to feedback by email at feedback@newscientist.com. Include your home address. This week and past feedback can be found on our website.

Source: www.newscientist.com

Harvard team makes significant strides in error correction technology

Posted on by
Reply

Quantum computing has advanced significantly with a new platform from Harvard University that is capable of dynamic reconfiguration and can demonstrate low error rates in two-qubit entangled gates. This breakthrough, highlighted in a recent Nature paper, represents a major advance in overcoming the challenges of quantum error correction and places Harvard’s technology alongside other leading quantum computing methods. Masu. This research, in collaboration with MIT and others, represents an important step toward scalable, error-correcting quantum computing. Credit: SciTechDaily.com

A method developed by a team at Harvard University to reduce errors addresses a critical hurdle in scaling up technology.

Quantum computing technology has the potential to achieve unprecedented speed and efficiency, vastly exceeding the capabilities of even the most advanced supercomputers currently available. However, this innovative technology has not been widely scaled or commercialized, primarily due to inherent limitations in error correction. Quantum computers, unlike classical computers, cannot correct errors by copying encoded data over and over again. Scientists had to find another way.

Now, a new paper Nature depicting Harvard University quantum computing A potential platform to solve a long-standing problem known as quantum error correction.

The Harvard team is led by quantum optics expert Mikhail Lukin, Joshua and Beth Friedman Professor of Physics and co-director of the Harvard Quantum Initiative. The research reported in Nature was a collaboration between Harvard University. Massachusetts Institute of Technology, Boston-based QuEra Computing. George Busmer Leverett Professor of Physics and Marcus Greiner’s group also participated.

Unique Harvard Platform

The Harvard University platform, an effort over the past several years, is built on an array of very cold rubidium atoms captured by a laser.Each atom They act as bits (called “qubits” in the quantum world) that can perform extremely fast calculations.

The team’s main innovation is configuring a “neutral atomic array” so that the layout can be dynamically changed by moving and connecting atoms during calculations. This is called “entanglement” in physics terms. 2 Operations that entangle pairs of atoms called qubit logic gates are units of computing power.

Running complex algorithms on a quantum computer requires many gates. However, these gating operations are known to be error-prone, and the accumulation of errors renders the algorithm useless.

In a new paper, the team reports near-perfect performance of the two-qubit entanglement gate with extremely low error rates. For the first time, they demonstrated the ability to entangle atoms with an error rate of less than 0.5 percent. In terms of operational quality, this puts the performance of the company’s technology on par with other major types of quantum computing platforms, such as superconducting qubits and trapped ion qubits.

Benefits and future prospects

However, Harvard’s approach has significant advantages over these competitors due to its large system size, efficient qubit control, and the ability to dynamically reconfigure the atomic layout.

“We demonstrate that the physical errors of this platform are low enough that we can actually imagine large-scale error correction devices based on neutral atoms,” said lead author and Harvard University Griffin School of Arts and Sciences. student Simon Evered said. group. “Currently, our error rates are low enough that if we group atoms into logical qubits (information is stored non-locally between the constituent atoms), we can Errors can be even lower than individual atoms.”

The Harvard team’s progress was tracked by former Harvard graduate student and current princeton university, former Harvard University postdoctoral fellow Manuel Endres, now at the California Institute of Technology. Taken together, these advances lay the foundation for quantum error correction algorithms and large-scale quantum computing. All of this means that quantum computing on neutral atomic arrays is reaching its full potential.

“These contributions open the door to very special opportunities in scalable quantum computing, and truly exciting times ahead for the field as a whole,” Lukin said.

Reference: “High-fidelity parallel entanglement gates on neutral atom quantum computers” Simon J. Evered, Dolev Bluvstein, Marcin Kalinowski, Sepehr Ebadi, Tom Manovitz, Hengyun Zhou, Sophie H. Li, Alexandra A. Geim, Tout T Wang, Nishad Maskara, Harry Levine, Julia Semeghini, Markus Greiner, Vladan Vretić, Mikhail D. Lukin, October 11, 2023. Nature.
DOI: 10.1038/s41586-023-06481-y

This research was supported by the U.S. Department of Energy’s Quantum Systems Accelerator Center. Ultracold Atom Center. National Science Foundation. Army Research Office Interdisciplinary University Research Initiative.And thatDARPAOptimization with a noisy intermediate-scale quantum device program.

Source: scitechdaily.com