How to Teach Robot to Laugh at the Right Time

Laughter comes in many forms, from a polite chuckle to a contagious howl of mirth. Scientists are now developing an AI system that aims to recreate these nuances of humour by laughing in the right way at the right time. The team behind the laughing robot, which is called Erica, say that the system could improve natural conversations between people and AI systems.

CLICK THE PHOTO TO ENJOY THE VIDEO

We think that one of the important functions of conversational AI is empathy,” said Dr Koji Inoue, of Kyoto University, the lead author of the research, published in Frontiers in Robotics and AI. “So we decided that one way a robot can empathise with users is to share their laughter.

 Inoue and his colleagues have set out to teach their AI system the art of conversational laughter. They gathered training data from more than 80 speed-dating dialogues between male university students and the robot, who was initially teleoperated by four female amateur actors.

The dialogue data was annotated for solo laughs, social laughs (where humour isn’t involved, such as in polite or embarrassed laughter) and laughter of mirth. This data was then used to train a machine learning system to decide whether to laugh, and to choose the appropriate type. It might feel socially awkward to mimic a small chuckle, but empathetic to join in with a hearty laugh. Based on the audio files, the algorithm learned the basic characteristics of social laughs, which tend to be more subdued, and mirthful laughs, with the aim of mirroring these in appropriate situations.

Our biggest challenge in this work was identifying the actual cases of shared laughter, which isn’t easy because as you know, most laughter is actually not shared at all,” said Inoue. “We had to carefully categorise exactly which laughs we could use for our analysis and not just assume that any laugh can be responded to.

The team tested out Erica’s “sense of humour” by creating four short dialogues for it to share with a person, integrating the new shared-laughter algorithm into existing conversation software. These were compared to scenarios where Erica didn’t laugh at all or emitted a social laugh every time she detected laughter.

The clips were played to 130 volunteers who rated the shared-laughter algorithm most favourably for empathy, naturalness, human-likeness and understanding. The team said laughter could help create robots with their own distinct character. “We think that they can show this through their conversational behaviours, such as laughing, eye gaze, gestures and speaking style,” said Inoue, although he added that it could take more than 20 years before it would be possible to have a “casual chat with a robot like we would with a friend.”

Source: https://www.theguardian.com/

How to Convert Carbon Dioxide To Oxygen on Mars

The growing list of “firsts” for Perseverance, NASA’s newest six-wheeled robot on the Martian surface, includes converting some of the Red Planet’s thin, carbon dioxide-rich atmosphere into oxygen. A toaster-size, experimental instrument aboard Perseverance called the Mars Oxygen In-Situ Resource Utilization Experiment (MOXIE) accomplished the task. The test took place April 20, the 60th Martian day, or sol, since the mission landed Feb. 18. While the technology demonstration is just getting started, it could pave the way for science fiction to become science fact – isolating and storing oxygen on Mars to help power rockets that could lift astronauts off the planet’s surface. Such devices also might one day provide breathable air for astronauts themselves. MOXIE is an exploration technology investigation – as is the Mars Environmental Dynamics Analyzer (MEDA) weather station – and is sponsored by NASA’s Space Technology Mission Directorate (STMD) and Human Exploration and Operations Mission Directorate.

This is a critical first step at converting carbon dioxide to oxygen on Mars,” said Jim Reuter, associate administrator for STMD. “MOXIE has more work to do, but the results from this technology demonstration are full of promise as we move toward our goal of one day seeing humans on Mars. Oxygen isn’t just the stuff we breathe. Rocket propellant depends on oxygen, and future explorers will depend on producing propellant on Mars to make the trip home.” For rockets or astronauts, oxygen is key, said MOXIE’s principal investigator, Michael Hecht of the Massachusetts Institute of Technology’s Haystack Observatory.

To burn its fuel, a rocket must have more oxygen by weight. Getting four astronauts off the Martian surface on a future mission would require approximately 15,000 pounds (7 metric tons) of rocket fuel and 55,000 pounds (25 metric tons) of oxygen. In contrast, astronauts living and working on Mars would require far less oxygen to breathe. “The astronauts who spend a year on the surface will maybe use one metric ton between them,” Hecht said.

Hauling 25 metric tons of oxygen from Earth to Mars would be an arduous task. Transporting a one-ton oxygen converter – a larger, more powerful descendant of MOXIE that could produce those 25 tons – would be far more economical and practicalMarsatmosphere is 96% carbon dioxide. MOXIE works by separating oxygen atoms from carbon dioxide molecules, which are made up of one carbon atom and two oxygen atoms. A waste product, carbon monoxide, is emitted into the Martian atmosphere. The conversion process requires high levels of heat to reach a temperature of approximately 1,470 degrees Fahrenheit (800 Celsius). To accommodate this, the MOXIE unit is made with heat-tolerant materials. These include 3D-printed nickel alloy parts, which heat and cool the gases flowing through it, and a lightweight aerogel that helps hold in the heat. A thin gold coating on the outside of MOXIE reflects infrared heat, keeping it from radiating outward and potentially damaging other parts of Perseverance.

Source: https://www.nasa.gov/

Robots With Living Human Skin

Shoji Takeuchi and colleagues from the Department of Mechano-Informatics and the Graduate School of Information Science and Technology at the University of Tokyo have developed a method for coating a robotic finger with living human skin. Their findings were published in the journal Matter. Scientists believe a new class of skin-covered robots could more effectively interact with their human counterparts.

There are three benefits to using living cells as a coating material for robots. First, by using the same skin material as humans, a more human-like appearance can be achieved. Second, the biological properties of cells can be used to provide robot skin with multimodal and multichannel sensing capabilities, self-repair capabilities, and metabolic capabilities that are difficult to achieve with artificial materials alone. Third, by using biological materials, robots can be made more environmentally friendly,” Takeuchi told Syfy Wire.

To get the skin onto the robotic appendage, scientists submerged it in a combination of collagen and human skin cells. Over time, the mixture attached itself to the finger, creating a first layer of skin. A second liquid containing keratinocyte cells — the dominant cells found in the epidermis — was then applied creating an outer layer. After a couple of weeks, the robotic finger had skin which was comparable in width to our own. Previous studies grew skin-like structures separately and later applied them to a synthetic surface. This new strategy has benefits over previous methods, in that it allows for the application of skin over uneven surfaces.

We found that we could adapt the skin to the curved 3D surface shape by culturing it on site, rather than making it elsewhere and attaching it to the surface. By installing an appropriate anchor structure, the entire surface could be covered,” Takeuchi said.

At present, the skin does not deliver any sensory information to the robot, but the team is working on incorporating a nervous system for just that purpose. The skin also doesn’t include any circulatory system for delivering nutrients to the tissue. As a result, it needed external assistance to acquire nutrients and for the removal of waste products. That means it spent a considerable portion of its time in a bath of sugars and amino acids.

“We are conceiving strategies to build circulatory systems within the skin. Another challenge is to develop more sophisticated skin with skin-specific functions by reproducing various organs in the skin such as sensory neurons, hair follicles, nails, and sweat glands,” Takeuchi explained.

That’s not to say the skin isn’t impressive even as it exists today. The current version was able to stretch with the finger as it bent or straightened and even healed itself after injury. Researchers made a small cut on the surface of the finger and then applied a collagen bandage. The cells of the skin then connected to the bandage and incorporated it into the skin, healing the wound.

Of course, the process will need to be scaled up if researchers hope to cover an entire humanoid robot in convincing human skin. A robot with disconnected pieces of skin might be even more terrifying to its human acquaintances than one with no skin at all. Now, that would be a dystopian nightmare better left to our fictions.

Source: https://www.u-tokyo.ac.jp/

Nano-Robots Injected into your Bloodstream to Fight Disease

What if there was a magical robot that could cure any disease? Don’t answer that. It’s a stupid question. Everyone knows there’s no one machine that could do that. But maybe a swarm made up of tens of thousands of tiny autonomous micro-bots could? That’s the premise laid out by proponents of nanobot medical technology. In science fiction, the big idea usually involves creating tiny metal robots via some sort of magic-adjacent miniaturization technology.

Luckily for us, the reality of nanobot tech is infinitely cooler. A team of researchers from Australia have developed a mind-blowing prototype that could work as a proof-of-concept for the future of medicine. Called “autonomous molecular machines,” the new nanotechnology eschews the traditional visage of microscopic metal automatons in favor of a more natural approach.

Inspired by biology, we design and synthesize a DNA origami receptor that exploits multivalent interactions to form stable complexes that are also capable of rapid subunit exchange”, explained the researchers. “DNA nanobots are synthetic nanometer-sized machines made of DNA and proteins. They’re autonomous because DNA itself is a self-assembling machine. Our natural DNA not only carries the code our biology is written in, it also knows when to execute. That’s part of the reason why, for example, your left and right feet tend to grow at roughly the same rate.”

Previous work in the field of DNA nanotechnology has demonstrated self-assembling machines capable of transferring DNA code, much like their natural counterparts. But the new tech out of Australia is unlike anything we’ve ever seen before.
Using the DNA origami receptor to demonstrate stable interactions with rapid exchange of both DNA and protein subunits, thus highlighting the applicability of the approach to arbitrary molecular cargo, an important distinction with canonical toehold exchange between single-stranded DNA. These particular nanobots can transfer more than just DNA information. Theoretically speaking, they could deliver any conceivable combination of proteins throughout a given biological system. To put that in simpler terms: the scientists should be able to eventually program swarms of these nanobots to hunt down bacteria, viruses, and cancer cells inside of our bodies. Each member of the swarm could carry a specific protein and, when they’ve found a bad cell, they could assemble their proteins into a formation designed to eliminate the threat.

Source: https://thenextweb.com/

Robot Performs much Better than Humans at Surgery

For years, the world of medicine has been steadily advancing the art of robot-assisted procedures, enabling doctors to enhance their technique inside the operating theatre. Now US researchers say a robot has successfully performed keyhole surgery on pigs all on its own without the guiding hand of a human. Furthermore, they add, the robot surgeon produced “significantly better” results than humans.

Smart Tissue Autonomous Robot (Star) carried out laparoscopic surgery to connect two ends of an intestine in four pigs. The robot excelled at the procedure, which requires a high level of precision and repetitive movements

Axel Krieger, of Johns Hopkins University, said it marked the first time a robot had performed laparoscopic surgery without human help. “Our findings show that we can automate one of the most intricate and delicate tasks in surgery: the reconnection of two ends of an intestine,” he said. “The Star performed the procedure in four animals and it produced significantly better results than humans performing the same procedure.”

Connecting two ends of an intestine is a challenging procedure in gastrointestinal surgery, requiring a surgeon to apply stitches – or sutures – with high accuracy and consistency. Even a slight hand tremor or misplaced stitch can result in a leak that could result in a patient suffering fatal complications. Krieger, an assistant professor of mechanical engineering at Johns Hopkins, helped create the robot, a vision-guided system designed specifically to suture soft tissue. It improves a 2016 model that repaired a pig’s intestines, but required a large incision to access the intestine and more guidance from humans.
Experts say new features allow for improved surgical precision, including specialised suturing tools and imaging systems that provide more accurate visualisations of the surgical field.
Source: https://www.theguardian.com/

Tesla Robot + Neuralink has Revolutionary Healthcare Applications

Elon Musk and his companies have a commitment to fearless innovation. The incredible accomplishments that his companies have achieved include cutting-edge electric vehicles with Tesla, next-generation space-flight capabilities with SpaceX, and the development of critical brain-machine interfaces with Neuralink, to name a few.

Musk’s most recent announcement was on behalf of Tesla, and was yet another ode to fearless innovation. Last week, during Tesla’s much anticipated “AI Day,” an event meant to showcase the company’s revolutionary strides in artificial intelligence technology, Musk announced the next frontier for the company: developing the Tesla Bot, a “general purpose, bi-pedal, humanoid robot capable of performing tasks that are unsafe, repetitive or boring.”

Elon Musk described the project in detail: “Basically, if you think about what we’re doing right now with the cars, Tesla is arguably the world’s biggest robotics company…because our cars are like semi-sentient robots on wheels. With the full self-driving computer, the inference engine on the car (which will keep evolving, obviously), Dojo, and all the neural nets recognizing the world, understanding how to navigate through the world, it kind of makes sense to put that onto a humanoid form.” Musk described the purpose behind this bot, atleast initially: “it’s intended to be friendly of course, and navigate through a world built for humans and eliminate dangerous repetitive and boring tasks.” Musk also explained that a useful humanoid robot should be able to navigate the world without being explicitly trained step-by-step, and instead, should be able to perform advanced tasks with cognitive understanding of simple commands, such as“pick up groceries.”

Source: https://www.forbes.com/

All-Terrain NanoRobot Flips Through A Live Colon

A rectangular robot as tiny as a few human hairs can travel throughout a colon by doing back flips, Purdue University engineers have demonstrated in live animal models. Why the back flips? Because the goal is to use these robots to transport drugs in humans, whose colons and other organs have rough terrain. Side flips work, too. Why a back-flipping robot to transport drugs? Getting a drug directly to its target site could remove side effects, such as hair loss or stomach bleeding, that the drug may otherwise cause by interacting with other organs along the way.

The study, published in the journal Micromachines, is the first demonstration of a microrobot tumbling through a biological system in vivo. Since it is too small to carry a battery, the microrobot is powered and wirelessly controlled from the outside by a magnetic field.


CLICK ON THE IMAGE TO ENJOY THE VIDEO

When we apply a rotating external magnetic field to these robots, they rotate just like a car tire would to go over rough terrain,” said David Cappelleri, a Purdue associate professor of mechanical engineering. “The magnetic field also safely penetrates different types of mediums, which is important for using these robots in the human body.

The researchers chose the colon for in vivo experiments because it has an easy point of entry – and it’s very messy. “Moving a robot around the colon is like using the people-walker at an airport to get to a terminal faster. Not only is the floor moving, but also the people around you,” said Luis Solorio, an assistant professor in Purdue’s Weldon School of Biomedical Engineering. “In the colon, you have all these fluids and materials that are following along the path, but the robot is moving in the opposite direction. It’s just not an easy voyage.

But this magnetic microrobot can successfully tumble throughout the colon despite these rough conditions, the researchers’ experiments showed. The team conducted the in vivo experiments in the colons of live mice under anesthesia, inserting the microrobot in a saline solution through the rectum. They used ultrasound equipment to observe in real time how well the microrobot moved around.

Source: https://www.purdue.edu/

How To Make Robots More Effective On The Future Battlefield

In an effort to make robots more effective and versatile teammates for Soldiers in combat, Army researchers are on a mission to understand the value of the molecular living functionality of muscle, and the fundamental mechanics that would need to be replicated in order to artificially achieve the capabilities arising from the proteins responsible for muscle contraction.

Bionanomotors, like myosins that move along actin networks, are responsible for most methods of motion in all life forms. Thus, the development of artificial nanomotors could be game-changing in the field of robotics research.

Researchers from the U.S. Army Combat Capabilities Development Command‘s Army Research Laboratory ‘(CCDC ARL) have been looking to identify a design that would allow the artificial nanomotor to take advantage of Brownian motion, the property of particles to agitatedly move simply because they are warm.

The CCDC ARL researchers believe understanding and developing these fundamental mechanics are a necessary foundational step toward making informed decisions on the viability of new directions in robotics involving the blending of synthetic biology, robotics, and dynamics and controls engineering.

Army researchers are on a mission to understand the value of the molecular ‘living’ functionality of muscle, and the fundamental mechanics that would need to be replicated in order to artificially achieve the capabilities arising from the proteins responsible for muscle contraction

By controlling the stiffness of different geometrical features of a simple lever-arm design, we found that we could use Brownian motion to make the nanomotor more capable of reaching desirable positions for creating linear motion,” said Dean Culver, a researcher in CCDC ARL’s Vehicle Technology Directorate. “This nano-scale feature translates to more energetically efficient actuation at a macro scale, meaning robots that can do more for the warfighter over a longer amount of time.”

These widely accepted muscle contraction models are akin to a black-box understanding of a car engine,” Culver explained. “More gas, more power. It weighs this much and takes up this much space. Combustion is involved. But, you can’t design a car engine with that kind of surface-level information. You need to understand how the pistons work, and how finely injection needs to be tuned. That’s a component-level understanding of the engine. We dive into the component-level mechanics of the built-up protein system and show the design and control value of living functionality as well as a clearer understanding of design parameters that would be key to synthetically reproducing such living functionality.”

Culver stated that the capacity for Brownian motion to kick a tethered particle from a disadvantageous elastic position to an advantageous one, in terms of energy production for a molecular motor, has been illustrated by ARL at a component level, a crucial step in the design of artificial nanomotors that offer the same performance capabilities as biological ones.

This research adds a key piece of the puzzle for fast, versatile robots that can perform autonomous tactical maneuver and reconnaissance functions,” Culver said. “These models will be integral to the design of distributed actuators that are silent, low thermal signature and efficient – features that will make these robots more impactful in the field.”

Culver noted that they are silent because the muscles don’t make a lot of noise when they actuate, especially compared to motors or servos, cold because the amount of heat generation in a muscle is far less than a comparable motor, and efficient because of the advantages of the distributed chemical energy model and potential escape via Brownian motion.

According to Culver, the breadth of applications for actuators inspired by the biomolecular machines in animal muscles is still unknown, but many of the existing application spaces have clear Army applications such as bio-inspired robotics, nanomachines and energy harvesting.

 

The Journal of Biomechanical Engineering recently featured their research.

Source: https://www.arl.army.mil/

How To Merge Your Brain With A.I.

Elon Musk said startup Neuralink, which aims to build a scalable implant to connect human brains with computers, has already implanted chips in rats and plans to test its brain-machine interface in humans within two years, with a long-term goal of peoplemerging with AI.” Brain-machine interfaces have been around for awhile. Some of the earliest success with the technology include Brown University’s BrainGate, which first enabled a paralyzed person to control a computer cursor in 2006. Since then a variety of research groups and companies, including the University of Pittsburgh Medical Center and DARPA-backed Synchron, have been working on similar devices. There are two basic approaches: You can do it invasively, creating an interface with an implant that directly touches the brain, or you can do it non-invasively, usually by electrodes placed near the skin. (The latter is the approach used by startup CTRL-Labs, for example.)

Neuralink, says Musk, is going to go the invasive route. It’s developed a chip containing an array of up to 96 small, polymer threads, each with up to 32 electrodes that can be implanted into the brain via robot and a 2 millimeter incision. The threads are small — less than 6 micrometers — because, as Musk noted in remarks delivered Tuesday night and webcast, Once implanted, according to Musk, the chip would connect wirelessly to devices. “It basically Bluetooths to your phone,” he said. “We’ll have to watch the App Store updates to that one,” he added (the audience laughed).

Musk cofounded Neuralink in 2017 and serves as the company’s CEO, though it’s unclear how much involvement he has given that he’s also serving as CEO for SpaceX and Tesla. Company cofounder and president, Max Hodak, has a biomedical engineering degree from Duke and has cofounded two other companies, MyFit and Transcriptic. Neuralink has raised $66.27 million in venture funding so far, according to Pitchbook, which estimates the startup’s valuation at $509.3 million. Both Musk and Hodak spoke about the potential for its company’s neural implants to improve the lives of people with brain damage and other brain disabilities. Its first goal, based on its discussions with such patients, is the ability to control a mobile device.

The company’s long-term goal is a bit more fantastical, and relates to Musk’s oft-repeated concerns over the dangers of advanced artificial intelligence. That goal is to use the company’s chips to create a “tertiary level” of the brain that would be linked to artificial intelligence.We can effectively have the option of merging with AI,” he said. “After solving a bunch of brain related diseases there is the mitigation of the existential threat of AI,” he continued.

CLICK ON THE IMAGE TO ENJOY THE VIDEO

In terms of progress, the company says that it has built a chip and a robot to implant it, which it has implanted into rats. According to the whitepaper the company has published (which has not yet undergone any peer review), it was able to record rat brain activity from its chips, and with many more channels than exist on current systems in use with humans. The first human clinical trials are expected for next year, though Hodak mentioned that the company has not yet begun to the FDA processes needed to conduct those tests.

Source: https://www.forbes.com/

Robot Farmer Operates 20 Hours A Day, Harvesting Tens Of Thousands Of Raspberries

Fieldwork Robotics, a University of Plymouth spin-off company, is developing an autonomous harvesting robot platform. A number of flexible robot arms attached to the platform will be able to pick raspberries, tomatoes, and other crops without crushing them or destroying the plant.

Fieldwork Robotics has completed initial field trials of its robot raspberry harvesting system. The tests took place at a West Sussex farm owned by Fieldwork’s industry partner, leading UK soft-fruit grower Hall Hunter Partnership, which supplies Marks & Spencer, Tesco and Waitrose. Data from the trials will be used to refine and improve the prototype system before further field trials are held later this year. If they are successful, manufacturing of a commercial system is expected to begin in 2020.

Fieldwork Robotics was incorporated to develop and commercialise the work of Dr Martin Stoelen, Lecturer in Robotics at the University’s School of Computing, Electronics and Mathematics.

CLICK ON THE IMAGE TO ENJOY THE VIDEO

Starting the field testing at Hall Hunter Partnership is a major milestone for us, and will give us invaluable feedback to keep developing the system towards commercialisation, as part of our Innovate UK funding. I am very proud of the achievements of the team, at Fieldwork Robotics Ltd and across my different research projects on robotic harvesting here at the University of Plymouth,  says Dr Martin Stoelen,

Farmers around the world are increasingly interested in robot technology to address the long-term structural decline in labourFieldwork is focusing initially on raspberries because they are hard to pick, are more delicate and easily damaged than other soft fruits, and grow on bushes with complex foliage and berry distribution. Once the system is proved to work with raspberries, it can be adapted readily for other soft fruits and vegetables, with the same researchers also developing proof-of-concept robots for other crops following interest from leading agribusinesses.

Source: https://www.plymouth.ac.uk/