A new artificial intelligence method removes the effect of gravity on cosmic images, showing the real shapes of distant galaxies.
Artificial intelligence (AI) seems to power all sorts of devices nowadays. Why not a lawn mower, too?
Jacopo Buongiorno and others say factory-built microreactors trucked to usage sites could be a safe, efficient option for decarbonizing electricity systems.
We may be on the brink of a new paradigm for nuclear power, a group of nuclear specialists suggested recently in The Bridge, the journal of the National Academy of Engineering. Much as large, expensive, and centralized computers gave way to the widely distributed PCs of today, a new generation of relatively tiny and inexpensive factory-built reactors, designed for autonomous plug-and-play operation similar to plugging in an oversized battery, is on the horizon, they say.
These proposed systems could provide heat for industrial processes or electricity for a military base or a neighborhood, run unattended for five to 10 years, and then be trucked back to the factory for refurbishment. The authors — Jacopo Buongiorno, MIT’s TEPCO Professor of Nuclear Science and Engineering; Robert Frida, a founder of GenH; Steven Aumeier of the Idaho National Laboratory; and Kevin Chilton, retired commander of the U.S. Strategic Command — have dubbed these small power plants “nuclear batteries.” Because of their simplicity of operation, they could play a significant role in decarbonizing the world’s electricity systems to avert catastrophic climate change, the researchers say. MIT News asked Buongiorno to describe his group’s proposal.
Circa 2019
Imagine we go through the disruptive transition between an economy where we need to work to make a living, to one where we don’t. It is hard to imagine because in North America; we haven’t been in this situation since the colonial era. Back in the colonial era, most people were farmers and families had to build their own homes. Neighbors traded with each other and with the closest town with what they had to get what else they needed. Those were difficult days with minimal supply chains established in North America. It is not a period we want to go back to, but we may learn from our forebears to prepare us for what is to come.
It is no surprise, in this age where automation is threatening to replace all employees, that we have concerns about how we can still function as a society when automation will take over most jobs. Fortunately, the same systems that threaten our livelihoods can bring us to a Golden Age of civilization where people live free, happy lives, without the concern for survival. I talk about the future of work in an article I published earlier this year. In a nutshell, and for the purpose of this article, I’ll jump to the conclusion: there won’t be enough demand for humans to have jobs within the next 20 years to sustain an employment-taxation type of economy.
Continue reading “A Fully Automated Economy–How Can It Work?” »
Sky surveys are invaluable for exploring the universe, allowing celestial objects to be catalogued and analyzed without the need for lengthy observations. But in providing a general map or image of a region of the sky, they are also one of the largest data generators in science, currently imaging tens of millions to billions of galaxies over the lifetime of an individual survey. In the near future, for example, the Vera C. Rubin Observatory in Chile will produce 20 TB of data per night, generate about 10 million alerts daily, and end with a final data set of 60 PB in size.
As a result, sky surveys have become increasingly labor-intensive when it comes to sifting through the gathered datasets to find the most relevant information or new discovery. In recent years machine learning has added a welcome twist to the process, primarily in the form of supervised and unsupervised algorithms used to train the computer models that mine the data. But these approaches present their own challenges; for example, supervised learning requires image labels that must be manually assigned, a task that is not only time-consuming but restrictive in scope; at present, only about 1% of all known galaxies have been assigned such labels.
Continue reading “Self-supervised machine learning adds depth, breadth and speed to sky surveys” »
DARPA announced the selection of four research teams to drive it home with no headlights on our Invisible Headlights program, which seeks to determine if it’s possible for autonomous vehicles to navigate in complete darkness using only passive sensors:
https://www.iflscience.com/plants-and-animals/uk-may-ban-boi…feel-pain/ More
DARPA has selected four industry and university research teams for the Invisible Headlights program, which seeks to determine if it’s possible for autonomous vehicles to navigate in complete darkness using only passive sensors.
Austrian shoe company Tec-Innovation has partnered with students at the Graz University of Technology in Austria to implement camera-based AI image recognition into their line of shoes that are specifically made to help those who are visually impaired.
The original version of these “seeing eye” shoes features ultrasonic sensors, which warn the person wearing them of obstacles in their way through haptic or auditory signals. AI image recognition that constantly learns, allows the shoes to provide more specific information to the wearer.
❤️ Check out the Gradient Dissent podcast by Weights & Biases: http://wandb.me/gd.
📝 The paper “Endless Loops: Detecting and Animating Periodic Patterns in Still Images” and the app are available here:
https://pub.res.lightricks.com/endless-loops/
Continue reading “This Magical AI Makes Your Photos Move! 🤳” »
These flaws in AI training give the technology a bad name, and so do regular media reports suggesting that intelligent machines are poised to decimate the human workforce. These themes, for many people, have obscured AI’s genuine usefulness in data analysis and conversational platforms. And while computer vision does indeed have its flaws, it is more than just a reflection of societal biases: it is potentially an essential tool for both society and business.
Computer vision, or CV, gives machines the power of visual recognition in a way that emulates human sight. Whether a machine is detecting dangers on the road or, more controversially, recognising faces in a crowd, the ultimate aim is to make decisions based on image interpretation.
The tech is an advanced form of pattern recognition, made through statistical comparison of data sets. This means that while machines can “see”, they have no real understanding of what they are looking at. They can distinguish one object from another, true, but can’t explain what this difference means.
These themes, for many people, have obscured AI’s genuine usefulness in data analysis and conversational platforms. And while computer vision does indeed have its flaws, it is more than just a reflection of societal biases: it is potentially an essential tool for both society and business.
Computer vision, or CV, gives machines the power of visual recognition in a way that emulates human sight. Whether a machine is detecting dangers on the road or, more controversially, recognising faces in a crowd, the ultimate aim is to make decisions based on image interpretation.
The tech is an advanced form of pattern recognition, made through statistical comparison of data sets. This means that while machines can “see”, they have no real understanding of what they are looking at. They can distinguish one object from another, true, but can’t explain what this difference means.
