Robots learn to shape letters using Play-Doh

https://www.engadget.com/robot-play-doh-letters-robocraft-202710608.html?src=rss

Humans aren’t the only ones working with Play-Doh. MIT CSAIL researchers have created a system, RoboCraft, that teaches robots how to work with the kid-friendly goo. The platform first takes the image of a shape (in this case, a letter of the alphabet) and reinterprets it as a cluster of interlocking particles. The bot then uses a physics-oriented neural network to predict how its two "fingers" can manipulate those spheres to match the desired outcome. A predictive algorithm helps the machine plan its actions.

The technology doesn’t require much time to produce usable results. It took just ten minutes of practice for an robot to perform roughly as well (and in some cases better) than humans remote-controlling the same hardware. That’s not the same as having a human shape the Play-Doh by hand, but it’s no mean feat for a machine discovering how to perform the task for the first time. Robots frequently struggle with soft objects where they tend to thrive with firm shapes.

RoboCraft-trained bots aren’t about to produce elaborate sculptures. The results are still imprecise, and the machine works slowly using just two fingers. The team is already developing a method of making dumplings, though, and plans to teach robots to use additional tools (such as a rolling pin) to prep the food.

The CSAIL scientists already have an idea of where the technology might be deployed. Kitchen robots could take over more responsibilities, while artistic automatons might create pottery. Eventually, technology like this could help the elderly and people with mobility issues by taking over household duties that require subtle motor skills.

via Engadget http://www.engadget.com

June 21, 2022 at 03:48PM

The Morning After: Cryptocurrency may be more centralized than you thought

https://www.engadget.com/the-morning-after-cryptocurrency-may-be-more-centralized-than-you-thought-111254399.html?src=rss

One of the boons of cryptocurrency is meant to be that no particular company, central bank or government has control. Er, right?

That might not be true. Researchers for a report commissioned by the Defense Advanced Research Projects Agency (DARPA) found there can be "unintended centralities" in these supposed decentralized systems.

Cryptocurrency power is concentrated among people or organizations with a large chunk of the pie. Almost like any other capitalist system? Gasp.

"Unintended centralities" was the term used, defined as circumstances where an entity has sway over a so-called decentralized system. This could give them the opportunity to tamper with records of ownership. The report also notes three ISPs handle 60 percent of all bitcoin traffic.

The report said 21 percent of nodes are running an old, vulnerable version of the core bitcoin client. Attackers could target these nodes and take over the majority of a blockchain network. Theoretically, at least. But there have been plenty of cryptocurrency attacks in the last few years. Nothing wrong with some skepticism.

Real-life examples already exist: Read CNBC’s report on lending platform Solend. It’s had issues with one major account holder wielding influence over the entire platform.

— Mat Smith

 

The biggest stories you might have missed

Proteus is Amazon’s first fully autonomous warehouse robot

Amazon has also introduced a new robotic arm at its fulfillment centers.

TMA
Amazon

It’s an imposing name for Amazon’s first autonomous warehouse robot, but it still looks like an industrial Roomba. Proteus can move around Amazon’s facilities on its own while carrying carts full of packages. The company said the robot uses an "advanced safety, perception and navigation technology," so it can do its work without getting in the way of human employees.

Continue reading.

Microsoft will phase out facial recognition AI that could detect emotions

The move comes as Microsoft pushes for more responsible uses of AI.

Microsoft will "retire" facial recognition technology it said could infer emotions as well as characteristics like age, gender and hair. The AI raised privacy questions, Microsoft said, and offering a framework created the potential for discrimination and other abuses. There was also no clear consensus on the definition of emotions.

Continue reading.

New Philips Hue smart lights include its first portable rechargeable smart lamp

Signify has also launched a new sunrise effect for Philips Hue lights.

TMA
Philips

Signify (the company responsible for Hue lighting) has introduced a bunch of new Philips Hue smart lighting products, including its first portable rechargeable lamp designed for both indoors and outdoors. The Philips Hue Go portable table lamp has a silicone grip so you can take it with you wherever you go. It can last for up to 48 hours on a single charge. The lamp will be available by the end of summer for $160 in the US and £130 in the UK. The company has new sunrise lighting effects, brighter downlights and even a new floor lamp. For all your smart lighting desires, read on.

Continue reading.

The best smart displays you can buy

Plus our advice on how to choose one.

TMA
Engadget

Smart displays are the second-wave devices born out of the success of the Amazon Echo, Google Home and other smart speakers. Adding visual and tactile components to what were once devices you could only bark orders at makes them more functional and intriguing. Amazon and Google dominate the space, and we’ve just updated our guide to buying a smart display, and we have opinions!

Continue reading.

NASA finally succeeds with its Artemis 1 wet launch test

It has fully fueled the Space Launch System for the first time.

NASA encountered a couple of issues while conducting the Artemis 1 wet dress rehearsal, but it still checked off a major milestone by the end of the test. The agency was able to fully fuel all the Space Launch System’s propellant tanks for the first time and proceed to terminal launch countdown. Wet dress rehearsals are tests that simulate a rocket launch without the rocket actually lifting off.

Continue reading.

via Engadget http://www.engadget.com

June 22, 2022 at 06:19AM

How Mathematics Solved The Burger Flipping Problem

https://www.discovermagazine.com/the-sciences/how-mathematics-solved-the-burger-flipping-problem


One of the sadly more neglected branches of research is the science of burger flipping. During the BBQ months, a key conundrum for the gastronomically inclined is whether a burger flipped many times cooks more quickly than one flipped only once. And if so, how many flips is ideal?

Today, we get an answer to this important question thanks to the work of Jean-Luc Thiffeault, a mathematician at the University of Wisconsin at Madison. Thiffeault has created a mathematical model of burger flipping which leads to a crucial insight that should allow you to cook burgers 29 per cent more quickly.

Thiffeault’s model is relatively simple. It assumes a flat burger that is one centimeter thick but infinite in extent (in mathematics, dreams can come true). The model includes a heating element on one side of the burger with fresh air on the other.

Heat enters the burger via direct contact with the heating element at 200 ⁰C, where the contact heat transfer coefficient is 900 Watts per square meter per degree centigrade, an apparently reasonable experimental value.

Escaping Heat

Heat escapes from the burger on the side exposed to the air at 25 ⁰C, where the radiation and heat transfer coefficient is 60 Watts per square meter per degree centigrade.

The meat is considered cooked when it reaches a temperature of 70 ⁰C so an important part of the calculation is to determine the temperature of the meat at the center of the patty. Note that the temperature history at each point in the burger is important. As Thiffeault points out, cooked meat can cool down but it cannot “uncook.”

All this allows Thiffeault to calculate how long it takes to cook a burger. He begins with the case when the burger is not flipped, a process he finds can take forever.

So what happens when the burger is flipped? One idea put forward by the famous science food author Harold McGee is that the faster a burger is flipped, the closer the cooking conditions come to a burger that is heated on both sides simultaneously. So flipping a burger many times, McGee suggested, should speed up the cooking process. But is this true?

Thiffeault puts this theory to the test by simulating the speed of cooking for different numbers of flips, while varying the interval between flips. He then uses the mathematics software MATLAB to optimize the process.

It turns out that for a single flip the optimal time for the first interval is just 45 per cent of the total and accounts for only 29 per cent of the cooking. The rest is done during the second interval, particularly towards the end when there is a sudden surge in the percentage of cooked meat. This is probably a result of the time it takes for heat from both intervals to diffuse through the meat.

This diffusion has important implications for greater number of flips. It turns out that for optimal cooking times, the flipping intervals should be about the same length, except for the last one, which should be longer to allow for diffusion.

But as the number of flips increases, the improvements in cooking time become smaller. Indeed, Thiffeault says they approach an asymptote suggesting that multiple flips can, at best, cook burgers 29 per cent more quickly than a single flip. “The improvement in optimal cooking time is fairly marginal after a few flips,” he says, dedicating his work to the mathematician Charlie Doering who died in 2021.

So the answer to the question “how many times should you flip a burger to optimize its cooking time?” is: a few times!

Flipping Improvements

For the many people who will flip just two or three times, Thiffeault offers one more piece of advice: “The shape of the cooking time function for two and three flips suggests that it is better to err on the side of cooking a bit longer for each flip, since a shorter interval leads to a longer cooking time.”

That’s interesting work but there are various ways the model could be improved. As it cooks, the burger’s moisture and fat content will change — a more advanced model might account for this.

Then there is the possibility of introducing a maximum temperature and how that would change the calculations. For many patty flippers, an unfortunate outcome to be avoided at all costs is a burger that is burnt on the outside but raw on the inside.

Does multiple flipping help? Probably. But to be sure, Thiffeault clearly has more flipping work to do.



Ref: The Mathematics of Burger Flipping : arxiv.org/abs/2206.13900

via Discover Main Feed https://ift.tt/7Hi0xj6

June 29, 2022 at 01:03PM