Science, Space, & Robotics - Page 225

The world continues marching towards the warm embrace of machines in the formation of Skynet, so it should come as no surprise that a new Defense Department report says the United States need to take "immediate action" to increase the development of its AI war technology.

The US military is behind in AI research and autonomous technology when compared against academic and private research, reports Engadget, which also adds that the US has been primarily focused on launching heavy, physical attacks like previous wars, that it has left itself behind the AI and autonomous technology race. Well, that needs to stop - and we can be sure that over the coming years US taxpayers' money will be spent on it, in a big way.

In the report, the Defense Department said that the Pentagon needs to gather intelligence on other nations' AI capabilities, and work on "counter-autonomy" solutions. So in other words, the bully didn't do his homework or bring lunch money, so he's going to cheat off your test and steal your lunch money and claim victory.

SpaceX has confirmed it has sent its next-generation rocket engine, codenamed Raptor, in for testing at a facility in Texas.

The new Raptor engine could be up to 3x as powerful as the current Merlin engines that power SpaceX's Falcon 9 and upcoming Falcon Heavy rockets, but details on the next-gen Raptor engine are thin right now. SpaceX President Gwynne Shotwell made the announcement of the testing of Raptor at the Small Satellite Conference in Utah.

SpaceX founder and real life Tony Stark, Elon Musk, has said that Raptor could have a thrust of around 500,000 pounds, which puts it in the same category as the main engine on a space shuttle. But unlike the shuttle, which uses three main engines and two booster rockets, the future Mars Colonial Transporter would be powered by nine Raptor engines, giving it plenty of power.

Google has applied for a patent on bathroom devices that monitor your health. Among them: a pressure and electrical-sensing mat, a color-sensing bathroom mirror, an ultrasonic bathtub, a pressure-sensing toilet seat, and a radar-field device.

Some monitor your body by sending out sound waves, electromagnetic waves, electrical signals, and various forms of radiation, among other techniques, the data from which could be sent to health professionals. The devices can monitor your blood pressure and various parts of your body to determine specific health issues, like an impaired knee or overworked heart.

This isn't Google's first foray into the medical technology field: previously it explored Google Lens, a device that monitored sugar levels via an electronic eye lens.

Google is getting all six degrees of separation with parent company Alphabet, and its life sciences company Verily - yeah, see - has announced it entered a new joint venture with GSK, one of the pharmaceutical powerhouses, called Galvani Bioelectronics.

Galvani Bioelectronics will be working on experimental medicine that uses will use electric signals and much smaller devices, in comparison to the current system of chemical drugs in order to treat chronic diseases. Verily, owned by Google along with GSK said they would invest over $700 million across 7 years, with GSK owning a 55% equity in the newly formed Galvani Bioelectronics. GSK is already familiar with bioelectronics, as it began work in the exciting new field back in 2012.

Verily has been working with much smaller medical devices, such as smart contact lens right after Google's X research labs formed. Verily has worked with another large part of big pharma, with a joint venture into medical robots with Johnson and Johnson. There have been reports from biotech insiders and former staff that have not agreed with Verily for working on too many 'far-fetched health care projects without focus", reports Recode.

In the future, all of our houses might be 3D printed - especially if it takes just 45 days, like the world's first 3D printed house that was recently build in China.

According to the reports, the entire house was 3D printed in one go, rather than being built with multiple pieces. Experts over at Huashang Tengda in China oversaw the project, which took 45 days to complete in Beijing's Tongzhou District.

After 45 days, what we have is a large 400-square-meter home, with the two-story villa featuring walls that are up to 8.2 feet thick. Each flooor has a height of 10 feet, so you can be super tall and not have to worry about getting around the house. The 3D printing side of things was completed by the company, with specialists overlooking the project, but there wasn't much physical work to be done as it did most of it on its own.

SpaceX has been on a roll lately with three consecutive successful rocket launches and landings. It all came to an end yesterday though when its Falcon 9 rocket failed to land on a floating drone ship at sea and promptly exploded (or suffered a "rapid unscheduled disassembly" as founder Elon Musk puts it), but not before it successfully launched two satellites into orbit.

Musk says the explosion was caused by low thrust in one of the rocket's three primary engines, and that the company is now working on upgrades to the rocket in order to handle the same situation in the future.

SpaceX has a carg resupply mission for the International Space Station next month, followed closely by a Falcon 9 ground landing in Cape Canaveral, Florida and a satellite launch in August.

China's current five-year economic plan came to light in March, and in it saw mention of the country's intent to build a manned deep-sea platform nearly 10,000 feet underwater. Now authorities have examined the particulars and decided to accelerate the project.

The purpose of the lab is multi-fold: to help hunt for treasure, evaluate mineral deposits (particularly oil), and to put it to military use (likely within the field of sensor and communication systems).

"Having this kind of long-term inhabited station has not been attempted this deep, but it is certainly possible," said Bryan Clark, a senior fellow at the Washington-based Center for Strategic and Budgetary Assessments. "Manned submersibles have gone to those depths for almost 50 years. The challenge is operating it for months at a time."

Our robot overlords are well and truly on their way, with former McDonald's USA CEO Ed Rensi, saying that it's better to buy $35,000 robots than it is to hire teenagers for $15 an hour.

During an interview with FOX Business Network's Mornings with Maria, Rensi said: "I was at the National Restaurant Show yesterday and if you look at the robotic devices that are coming into the restaurant industry -- it's cheaper to buy a $35,000 robotic arm than it is to hire an employee who's inefficient making $15 an hour bagging French fries -- it's nonsense and it's very destructive and it's inflationary and it's going to cause a job loss across this country like you're not going to believe".

He added: "It's not just going to be in the fast food business. Franchising is the best business model in the United States. It's dependent on people that have low job skills that have to grow. Well if you can't get people a reasonable wage, you're going to get machines to do the work. It's just common sense. It's going to happen whether you like it or not. And the more you push this it's going to happen faster".

China seems to have beat us to the punch when it comes to enlisting robots and AI to help fight crime. It's not quite on the level of Robo Cop, but the autonomous machine is still slightly deadly.

The researchers at CERN have just released an insane 300TB of data from the Large Hadron Collider, with the data itself from 2011.

A physicist who works on the Compact Muon Solenoid detector, Kati Lassila-Perini explains: "Once we've exhausted our exploration of the data, we see no reason not to make them available publicly. The benefits are numerous, from inspiring high school students to the training of the particle physicists of tomorrow. And personally, as CMS's data preservation coordinator, this is a crucial part of ensuring the long-term availability of our research data".

The raw data from the detectors, as well as the "derived" data sets can be used with tools released by CERN. There's even an entire CERN Linux environment, where you can boot up a virtual machine and start playing with scripts and apps.

DJI has been making some seriously good professional drones over the years, with its impressive Phantom series, but has just unveiled its latest is the Matrice 600, a hexacopter that is capable of adjusting how it flies automatically, depending on what camera is attached.

The M600 costs $4599, and is the latest drone in DJI's huge professional lineup that is a successor to its current "Spreading Wings" series, which are higher-end craft that include retractable landing gear, and a foldable design. The new M600 features dust-proof propellers, as well as self-cooling motors.

DJI's new M600 launches with an improved, robust A3 flight controller that will change flight parameters depending on what it's carrying, and the amazing Lightbridge 2 camera link. Lightbridge 2 delivers higher frame rates over 1080p live-streamed video back to the pilot, at up to 3 miles away. This video is pumped to the pilot at 1080p 60FPS, which is impressive given the 3-mile distance allowance.

We all know the future is filled with robots, so it should come as no surprise that the University of Science and Technology of China is showing off its impressively realistic robot, Jia Jia.

Jia Jia looks more human than previous robots, and is capable of interacting with real humans, and can make realistic facial impressions. Jia Jia can tell you if she senses that you're taking unflattering pictures of her, where she'll say: "Don't come too close to me when you are taking a picture. It will make my face look fat".

The researchers spent three years designing Jia Jia to make sure that her mouth moves when she speaks, and that her eyes glance around the room naturally. Jia Jia can't laugh or cry just yet, and her hands still don't look super realistic, yet. The next version of Jia Jia will look better, with the researchers continuing to work on the robot without any plans of mass production. Team leader Chen Xiaoping said they hope to give her deep learning and facial recognition in the near future.

I didn't think we'd see this for at least another 10 years, but the world's first cyborg Olympics (or Cybathlon) will happen in Zurich in October, this year. The event was created as a way of creating innovation in the industry, where only a few eligible for prostheses actually use them. A trial event took place last year, and was a success.

The Cybathlon will measure the performance of the latest developments in technology that assist people with disabilities in everyday tasks. The modifications done to humans are encouraged at the Cybathlon, compared to the normal Olympics where athletes with enhancements are considered to have an unfair advantage - and as I write this, it feels like I'm explaining some future DLC of Deus Ex: Mankind Divided.

The Swiss Federal Institute of Technology in Zurich (ETH Zurich) is organizing the event where in a cycling race (as an example) paraplegics will be pushed forward using electrical stimulation systems that will move their legs by stimulating the muscles. If competitors have prosthetic arms, contests will see them slicing loaves of bread and opening jars of jam. Other events will include people climbing up and down stairs, or walking across stepping stones.

Skynet has begun its takeover, of the library. Future University of Hakodate researchers have announced that their artificial intelligence has co-written a short-form novel, and it's been accepted by a Japanese story competition.

The short-form novel co-written by the AI has been accepted by the Hoshi Shinichi Literary Award, and while the story didn't win the competition, its acceptance is a huge win for AI systems becoming more capable of reaching human-like creativity.

The team was led by computer science professor Hitoshi Matsubara, who worked closely with their AI during the writing process. The team assigned a gender to the protagonist in the novel, and then developed a rough outline of the plot. The team created a list of words, phrases, and sentences to be included with the story. The AI has the job of assembling the distinct assets into a unified text that was just intelligible, but compelling, as well.

Yesterday Microsoft launched its teen girl 'Tay' Twitter AI. It's designed to interact with followers and become smarter for it, so naturally it was doomed to fail. Within 24 hours, the Internet transformed it into a pro-Hitler sex robot, as the Internet is wont to do.

The tweets in question are deleted now, but screencaps like those below give you a healthy idea of what occurred. Warning: they're not for the easily offended.

Since this happened, Tay temporarily went offline, so Microsoft could "make adjustments." A cursory look through her current feed shows nothing out of the ordinary, so the situation is under control...for now.

Google parent company Alphabet Inc. have decided its robotics outfit Boston Dynamics isn't likely to produce marketable products anytime soon and as such, have put it up for sale, two sources familiar with the matter say.

The rumor is in line with accidentally leaked e-mails which indicate Google's concern with the viability of the company. The public reaction to the Atlas robot revealed last month could be a factor in the decision as well; the e-mails do well to show Google's worry.

"There's excitement from the tech press, but we're also starting to see some negative threads about it being terrifying, ready to take humans' jobs," wrote Courtney Hohne, a director of communications at Google and the spokeswoman for advanced research group Google X.

The AlphaGo AI created by Google's DeepMind division yesterday lost its first Go match to world champion Lee Sedol in a five-game series. Commentator Song Taegon remarked that Sedol had developed a better understanding of his opponent, making for a highly anticipated game 5. Sedol was beaten again, although it seems to have been a much tighter game, with commentators going back and forth right until the end about who was on top.

"It was difficult to say at what point AlphaGo was ahead or behind," said Michael Redmond, 9-dan, American commentator. "A close game throughout. AlphaGo made what looked like a mistake with move 48, similar to the mistake in Game Four in the middle of the board. After that AlphaGo played very well in the middle of the board, and the game developed into a long, very difficult end game."

Microsoft is turning to a very interesting platform to help improve research in artificial intelligence. Minecraft is being used by researchers at Microsoft, and by select academic researchers through a private beta program, in its unstructured mode, as a sort of testing ground for testing AI theories and programming.

Now how would they go about doing that? The AIX platform which is being developed by a Microsoft lab in Cambridge, U.K., that they're using can potentially be programmed to learn in a similar and, well, general way that humans do, to program the Minecraft avatar to go forth and do stuff. It's a closed environment where researchers can look easily observe what the character is doing, and in a safe environment. The constructible world allows for a number of different variables that let them test in a lot of different conditions. All without actually sacrificing real, and expensive, robots.

"Minecraft is the perfect platform for this kind of research because it's this very open world," Katja Hofmann, the platforms developer, said. "You can do survival mode, you can do 'build battles' with your friends, you can do courses, you can implement our own games. This is really exciting for artificial intelligence because it allows us to create games that stretch beyond current abilities." Even trying to get your character to do simple tasks, like walking, is very beneficial to AI research.

After dominating the five-game Go tournament series 3-0 and looking invincible in the process, Google's DeepMind AI has finally fallen to its human opponent; world Go champion Lee Sedol made the series 3-1 over the weekend after seeing weaknesses in Deepmind's game and mounting a surprising comeback. The match was described as "long and complicated."

The fifth and final match will begin late tonight or early tomorrow morning depending where you live, and can be viewed here. It's highly anticipated, and for good reason.

"It seems Lee Sedol can now read AlphaGo better and has a better understanding of how AlphaGo moves," said Song Taegon, 9-dan, Korean commentator. "For the fifth match, it will be a far closer battle than before since [they] know each better."

Yesterday, Google's DeepMind AI took the first game against Go world champion Lee Sedol, to the shock of many. You might think it's just one game and Sedol could just as well come back to trounce 4-1, but today DeepMind did it again, putting itself up 2-0. Sedol, who after yesterday's loss commented he was still very confident he could beat the AI, is now very much on the backfoot.

"Yesterday I was surprised but today it's more than that - I am speechless," Lee remarked after the game. "I admit that it was a very clear loss on my part. From the very beginning of the game I did not feel like there was a point that I was leading."

DeepMind founder Demis Hassabis said the AI was confident in victory from the midway point onward.