Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 1,449

A mini brain with trillions of petaflops in your pant pocket? Sounds Good!

“This is what we’re announcing today,” said Knowles. “A machine that in fact will exceed the parametric capacity of the human brain.”

That next-gen IPU, he said, would realize the vision of 1960s compute scientist Jack Good, a colleague of Alan Turing’s who conceived of an “intelligence explosion.”

Those synapses are “very similar to the parameters that are learned by an artificial neural network.” Today’s neural nets have gotten close to a trillion, he noted, “so we clearly have another two or more orders of magnitude to go before we have managed to build an artificial neural network that has similar parametric capacity to a human brain.

The company said it is working on a computer design, called The Good Computer, which will be capable of handling neural network models that employ 500 trillion parameters, making possible what it terms super-human ultra-intelligence.

The Bow is the first chip to use what’s called wafer-on-wafer chip technology, where two die are bound together. It was developed in close collaboration with contract chip manufacturing giant Taiwan Semiconductor Manufacturing.

The chip can perform 350 trillion floating point per second of mixed-precision AI arithmetic, said Knowles, which he said made the chip the highest-performing AI processor in the world today.

Continue reading “AI computer maker Graphcore unveils 3D chip, promises 500-trillion-parameter ‘ultra-intelligence’ machine” | >

It’s easy to see why: as shockingly powerful mini-processors, neurons and their connections—together dubbed the connectome—hold the secret to highly efficient and flexible computation. Nestled inside the brain’s wiring diagrams are the keys to consciousness, memories, and emotion. To connectomics, mapping the brain isn’t just an academic exercise to better understand ourselves—it could lead to more efficient AI that thinks like us.

But often ignored are the brain’s supporting characters: astrocytes—brain cells shaped like stars—and microglia, specialized immune cells. Previously considered “wallflowers,” these cells nurture neurons and fine-tune their connections, ultimately shaping the connectome. Without this long-forgotten half, the brain wouldn’t be the computing wizard we strive to imitate with machines.

In a stunning new brain map published in Cell, these cells are finally having their time in the spotlight. Led by Dr. H. Sebastian Seung at Princeton University, the original prophet of the connectome, the map captures a tiny chunk of the mouse’s visual cortex, less than 1,000 times smaller than a pea. Yet jam-packed inside the map aren’t just neurons; in a technical tour de force, the team mapped all brain cells, their connections, blood vessels, and even the compartments inside cells that house DNA and produce energy.

Read more | >

It should come as little surprise that pioneering work in biological robotics is as controversial as it is exciting. Take for example the article published in December 2021 in the Proceedings of the National Academy of Sciences by Sam Kreigman and Douglas Blackiston at Tufts University and colleagues. This article, entitled “Kinematic self-replication in reconfigurable organisms,” is the third installment of the authors’ “xenobots” series.

Read more | >

“There are a lot of partnerships where you’re carpooling to go to the nightclub, and then there are long-term collaborations,” said Saubestre. “We are fiancés for the time being, but we are trying to make this couple work.”

“There’s a lot that can be optimized, and there’s a lot of potential, and that was the reason to go into this partnership with someone who is as keen as we are to make it happen,” said Saubestre. “We’re hoping to retain this competitive advantage that we’ve had over the years and inventing the future machines to run our simulations.”

Added Saubestre, some vendors with whom TotalEnergies works focus on selling cloud services, but such services aren’t necessarily ideal for the kinds of work the energy giant needs to get done.

In a paper published on February 23, 2022 in Nature Machine Intelligence, a team of scientists at the Max Planck Institute for Intelligent Systems (MPI-IS) introduce a robust soft haptic sensor named “Insight” that uses computer vision and a deep neural network to accurately estimate where objects come into contact with the sensor and how large the applied forces are. The research project is a significant step toward robots being able to feel their environment as accurately as humans and animals. Like its natural counterpart, the fingertip sensor is very sensitive, robust, and high-resolution.

The thumb-shaped sensor is made of a soft shell built around a lightweight stiff skeleton. This skeleton holds up the structure much like bones stabilize the soft finger tissue. The shell is made from an elastomer mixed with dark but reflective aluminum flakes, resulting in an opaque grayish color that prevents any external light finding its way in. Hidden inside this finger-sized cap is a tiny 160-degree fish-eye camera, which records colorful images, illuminated by a ring of LEDs.

When any objects touch the sensor’s shell, the appearance of the color pattern inside the sensor changes. The camera records images many times per second and feeds a deep neural network with this data. The algorithm detects even the smallest change in light in each pixel. Within a fraction of a second, the trained machine-learning model can map out where exactly the finger is contacting an object, determine how strong the forces are, and indicate the force direction. The model infers what scientists call a force map: It provides a force vector for every point in the three-dimensional fingertip.

Read more | >

Identifying a malfunction in the nation’s power grid can be like trying to find a needle in an enormous haystack. Hundreds of thousands of interrelated sensors spread across the U.S. capture data on electric current, voltage, and other critical information in real time, often taking multiple recordings per second.

Researchers at the MIT-IBM Watson AI Lab have devised a computationally efficient method that can automatically pinpoint anomalies in those in real time. They demonstrated that their artificial intelligence method, which learns to model the interconnectedness of the power grid, is much better at detecting these glitches than some other popular techniques.

Because the they developed does not require annotated data on power grid anomalies for training, it would be easier to apply in real-world situations where high-quality labeled datasets are often hard to come by. The model is also flexible and can be applied to other situations where a vast number of interconnected sensors collect and report data, like traffic monitoring systems. It could, for example, identify traffic bottlenecks or reveal how traffic jams cascade.

Read more | >

A new study challenges the conventional approach to designing soft robotics and a class of materials called metamaterials by utilizing the power of computer algorithms. Researchers from the University of Illinois Urbana-Champaign and Technical University of Denmark can now build multimaterial structures without dependence on human intuition or trial-and-error to produce highly efficient actuators and energy absorbers that mimic designs found in nature.

The study, led by Illinois civil and environmental engineering professor Shelly Zhang, uses optimization theory and an -based design process called . Also known as digital synthesis, the builds composite structures that can precisely achieve complex prescribed mechanical responses.

The study results are published in the Proceedings of the National Academy of Sciences.

Read more | >

Additive manufacturing, or 3D printing, can create custom parts for electromagnetic devices on-demand and at a low cost. These devices are highly sensitive, and each component requires precise fabrication. Until recently, though, the only way to diagnose printing errors was to make, measure and test a device or to use in-line simulation, both of which are computationally expensive and inefficient.

To remedy this, a research team co-led by Penn State created a first-of-its-kind methodology for diagnosing errors with machine learning in real time. The researchers describe this framework—published in Additive Manufacturing —as a critical first step toward correcting 3D-printing errors in real time. According to the researchers, this could make printing for sensitive devices much more effective in terms of time, cost and computational bandwidth.

“A lot of things can go wrong during the process for any component,” said Greg Huff, associate professor of electrical engineering at Penn State. “And in the world of electromagnetics, where dimensions are based on wavelengths rather than regular units of measure, any small defect can really contribute to large-scale system failures or degraded operations. If 3D printing a household item is like tuning a tuba—which can be done with broad adjustments—3D-printing devices functioning in the electromagnetic domain is like tuning a violin: Small adjustments really matter.”

Read more | >

Engineers sometimes turn to nature for inspiration. Cold Spring Harbor Laboratory Associate Professor Saket Navlakha and research scientist Jonathan Suen have found that adjustment algorithms—the same feedback control process by which the Internet optimizes data traffic—are used by several natural systems to sense and stabilize behavior, including ant colonies, cells, and neurons.

Internet engineers route data around the world in small packets, which are analogous to . As Navlakha explains, “The goal of this work was to bring together ideas from and Internet design and relate them to the way forage.”

The same algorithm used by internet engineers is used by ants when they forage for food. At first, the colony may send out a single ant. When the ant returns, it provides information about how much food it got and how long it took to get it. The colony would then send out two ants. If they return with food, the colony may send out three, then four, five, and so on. But if ten ants are sent out and most do not return, then the colony does not decrease the number it sends to nine. Instead, it cuts the number by a large amount, a multiple (say half) of what it sent before: only five ants. In other words, the number of ants slowly adds up when the signals are positive, but is cut dramatically lower when the information is negative. Navlakha and Suen note that the system works even if individual ants get lost and parallels a particular type of “additive-increase/multiplicative-decrease algorithm” used on the internet.

Read more | >