Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 806

Watch ‘world’s most powerful’ humanoid robot withstand brutal kicks

The Chinese firm, Unitree, claims that its upgraded humanoid robot, “powertrain provides the highest level of speed, power, maneuverability and flexibility.”

Chinese robotic systems firm Unitree marks a groundbreaking development with the upgrade of its humanoid robot.

The robot, called H1, has also been billed as the ‘world’s most powerful general-purpose humanoid robot’ with its advanced “powertrain [which] provides the highest level of speed, power, maneuverability and flexibility,” claims Unitree’s website.

Situated in Hangzhou, just outside Shanghai, Unitree Robotics was established in 2017. The company’s mission is to democratize legged robotics, aspiring to make them as widespread and cost-effective as smartphones and drones are in contemporary times.

Kiwi Navy will test new AI, solar-powered robot boat for endless recon

The Royal New Zealand Navy is currently awaiting the arrival of its latest Uncrewed Surface Vessel, the wind-powered “Bluebottle,” ahead of a 7-month sea trial.

The Royal New Zealand Navy (RNZN) will soon receive its first 22.3-foot (6.8-meter) long renewable-powered Uncrewed Surface Vessel (USV) to trial on a short-term lease, the New Zealand Defense Force has announced. Called “Bluebottle,” the USV will provide persistent surveillance around the waters of New Zealand for fishery protection, border protection, or meteorological data.

Autonomous border control

HMNZS Aotearoa is currently transporting the USV from Sydney to Auckland. Once operational, “Bluebottle” will undertake maritime tasks without fuel or personnel on its planned seven-month-long trial. Designed and built by Sydney-based Ocius Technology, the company has sold several USVs to the Australian Defence Force and collaborated with the Australian Border Force, energy, and scientific agencies.

Foretellix raises $85M to build and test scenarios for self-driving systems

On their way to building fully autonomous vehicles, self-driving car makers are facing a tall task: training their AIs to be able to respond reliably to any and all scenarios that a car, truck or bus might encounter as well as, or hopefully better, than a human would. Today, a startup with a platform to help with that challenge is announcing a sizeable round of funding to take those strategies up a gear.

Foretellix, which builds verification and validation solutions to test the full range of driver assistance and autonomous systems that are coming out on the market, has closed its Series C at $85 million. The round includes financial investors alongside strategic backers from the automotive and chip industries, a signal of who is already doing business with Foretellix, as well as the longer business trajectory for the startup.

The round is being led by Israeli VC 83North, with Singapore’s Temasek and carmaker Isuzu investing alongside Woven Capital (Toyota’s venture fund), Nvidia, Artofin and previous backers MoreTech, Nationwide, Volvo Group VC, Jump Capital, Next Gear Ventures and OurCrowd. Foretellix may ring a bell for readers: The first close of this round was in May of this year (at $43 million).

AI approach offers solutions to tricky optimization problems, from global package routing to power grid operation

While Santa Claus may have a magical sleigh and nine plucky reindeer to help him deliver presents, for companies like FedEx, the optimization problem of efficiently routing holiday packages is so complicated that they often employ specialized software to find a solution.

This software, called a mixed-integer linear programming (MILP) solver, splits a massive optimization problem into and uses generic algorithms to try and find the best solution. However, the solver could take hours—or even days—to arrive at a solution.

The process is so onerous that a company often must stop the software partway through, accepting a solution that is not ideal but the best that could be generated in a set amount of time.

Meta, IBM launch alliance to keep AI’s future open

Meta, IBM and dozens of startups and researchers have launched an alliance defending a more open and collaborative method to develop artificial intelligence, setting up a clash with OpenAI and Google over the technology’s future.

The philosophical debate has become the central battleground for AI’s future, with increasing concern that Microsoft-backed OpenAI and Google will alone underpin a technology that could become increasingly crucial to our everyday lives.

“This is a pivotal moment in defining the future of AI,” said IBM CEO Arvind Krishna in the statement announcing the AI Alliance on Tuesday.

Laser additive manufacturing: Listening for defects as they happen

Researchers from EPFL have resolved a long-standing debate surrounding laser additive manufacturing processes with a pioneering approach to defect detection.

The progression of laser additive —which involves 3D printing of metallic objects using powders and lasers—has often been hindered by unexpected defects. Traditional monitoring methods, such as and machine learning algorithms, have shown significant limitations. They often either overlook defects or misinterpret them, making precision manufacturing elusive and barring the technique from essential industries like aeronautics and automotive manufacturing.

But what if it were possible to detect defects in real-time based on the differences in the sound the printer makes during a flawless print and one with irregularities? Up until now, the prospect of detecting these defects this way was deemed unreliable. However, researchers at the Laboratory of Thermomechanical Metallurgy (LMTM) at EPFL’s School of Engineering have successfully challenged this assumption.

Enhanced AI tracks neurons in moving animals

Recent advances allow imaging of neurons inside freely moving animals. However, to decode circuit activity, these imaged neurons must be computationally identified and tracked. This becomes particularly challenging when the brain itself moves and deforms inside an organism’s flexible body, e.g., in a worm. Until now, the scientific community has lacked the tools to address the problem.

Now, a team of scientists from EPFL and Harvard have developed a pioneering AI method to track inside moving and deforming animals. The study, now published in Nature Methods, was led by Sahand Jamal Rahi at EPFL’s School of Basic Sciences.

The new method is based on a (CNN), which is a type of AI that has been trained to recognize and understand patterns in images. This involves a process called “convolution,” which looks at small parts of the picture—like edges, colors, or shapes—at a time and then combines all that information together to make sense of it and to identify objects or patterns.

The Threat of Advanced AI: Urgent Attention Needed

Geoffrey Miller is a professor of evolutionary psychology at the University of New Mexico, a researcher and an author.

Artificial Intelligence possesses the capability to process information thousands of times faster than humans. It’s opened up massive possibilities. But it’s also opened up huge debate about the safety of creating a machine which is more intelligent and powerful than we are. Just how legitimate are the concerns about the future of AI?

/* */