Lifeboat Foundation

Picking up a box and placing it in a neat pile is not an impressive action in itself for a robot; understanding an enigmatic human command and correctly deciphering and explaining the decision-making process, are definitely innovations. Digit owes parts of its progress to the generative artificial intelligence revolution that also reached the field of robotics and turned expectations from it on its head. “I’ve been asked what’s the biggest thing in 2024 besides language modeling — it’s robotics. Period,” Nvidia’s senior AI scientist Jim Fan wrote in December. “We’re about three years away from a ChatGPT moment for physical AI agents,” he explained.

Ever since Fan made this statement, it seems that everyone is talking about the “ChatGPT moment of robotics”, or the hope of a technological breakthrough that will push the field forward and finally fill our homes with intelligent humanoid robots to help us with household chores, wash the floor, set the table or do the laundry (but not fold it). “What has been happening in recent months is dramatic,” explains Amir Bousani, CEO of R-Go Robotics, which recently entered into a partnership with Nvidia to equip the robot it is developing with its spatial perception capabilities. “The physical world is more difficult than the internet,” notes Dr. Oren Etzioni, founding CEO of the Allen Institute for Artificial Intelligence, “but the field of robots that have the ability to behave in general is running much faster today.”

The huge interest in humanoid robots, or humanoids, which Fan is talking about, is evident in the constant announcements in the field: in February, the startup Figure raised $675 million from Jeff Bezos, Nvidia, and OpenAI for the development of the humanoid. In March, Nvidia’s CEO stood on the stage of the company’s developer conference alongside nine humanoids from different companies and announced that building models for robots is “one of the most exciting problems to solve in artificial intelligence”; in April, Elon Musk promised that he would launch the humanoid robot he is developing — Optimus — next year and predicted that by 2040 there will be a billion humanoids among us. A short time later, the activities of Mentee Robotics, Amnon Shashua’s company that was founded two years ago and also develops the humanoid, went public.

“My main life work, along with basic science, has been building our Wolfram language computational language for the purpose of having a way to express things computationally that’s useful to both humans and computers,” Wolfram told TechCrunch.

As AI developers and others start to think more deeply about how computers and people intersect, Wolfram says it is becoming much more of a philosophical exercise, involving thinking in the pure sense about the implications this kind of technology may have on humanity. That kind of complex thinking is linked to classical philosophy.

Continue reading “Stephen Wolfram thinks we need philosophers working on big questions around AI” »

The wing dynamics of flying animal species have been the inspiration for numerous flying robotic systems. While birds and bats typically flap their wings using the force produced by their pectoral and wing muscles, the processes underlying the wing movements of many insects remain poorly understood.

Researchers at Ecole Polytechnique Fédérale de Lausanne (EPFL, Switzerland) and Konkuk University (South Korea) recently set out to explore how herbivorous insects known as rhinoceros beetles deploy and retract their wings. The insight they gathered, outlined in a paper published in Nature, was then used to develop a new flapping microrobot that can passively deploy and retract its wings, without the need for extensive actuators.

“Insects, including beetles, are theoretically believed to use thoracic muscles to actively deploy and retract their wings at the wing bases, similarly to birds and bats,” Hoang-Vu Phan, the lead author of the paper, told Tech Xplore. “However, methods of recording or monitoring muscular activity still cannot determine which muscles beetles use to deploy and retract their wings nor explain how they do so.”

Over the past couple of decades, computer scientists have developed a wide range of deep neural networks (DNNs) designed to tackle various real-world tasks. While some of these models have proved to be highly effective, some studies found that they can be unfair, meaning that their performance may vary based on the data they were trained on and even the hardware platforms they were deployed on.

For instance, some studies showed that commercially available deep learning–based tools for facial recognition were significantly better at recognizing the features of fair-skinned individuals compared to dark-skinned individuals. These observed variations in the performance of AI, in great part due to disparities in the training data available, have inspired efforts aimed at improving the fairness of existing models.

Researchers at University of Notre Dame recently set out to investigate how hardware systems can contribute to the fairness of AI. Their paper, published in Nature Electronics, identifies ways in which emerging hardware designs, such as computing-in-memory (CiM) devices, can affect the fairness of DNNs.

The trailer featured quotes from famous film critics panning Francis Ford Coppola’s previous films that appear to be made up by ChatGPT.

Organoids, $500/month, last up to 100 days, used by select universities:

Designed by FinalSpark, biocomputers are a more efficient and low-energy alternative for training AI models.

Researchers at Argonne have developed an innovative technique that creates “fingerprints” of different materials that can be read and analyzed by a neural network to yield previously inaccessible information — https://bit.ly/3LCklZw.

The goal of the AI is just to treat the scattering patterns as…

Study shows how materials change as they are stressed and relaxed.

Continue reading “Scientists develop new artificial intelligence method to create material ‘fingerprints’” »

Have you ever wonder how SNNs work and their difference from traditional neural networks? Or how SNNs play an important role in computing beyond the Moore’s Law?

What is SNN?
Spiking neural network (SNN) is a new form of neural networks with biologically realistic mechanisms designed to emulate the efficiency and effectiveness of the biological brain.

Researchers worldwide can now create highly realistic brain cortical organoids — essentially miniature artificial brains with functioning neural networks — thanks to a proprietary protocol released this month by researchers at the University of California San Diego.

The new technique, published in Nature Protocols (“Generation of ‘semi-guided’ cortical organoids with complex neural oscillations”), paves the way for scientists to perform more advanced research regarding autism, schizophrenia and other neurological disorders in which the brain’s structure is usually typical, but electrical activity is altered. That’s according to Alysson Muotri, Ph.D., corresponding author and director of the UC San Diego Sanford Stem Cell Institute (SSCI) Integrated Space Stem Cell Orbital Research Center. The SSCI is directed by Dr. Catriona Jamieson, a leading physician-scientist in cancer stem cell biology whose research explores the fundamental question of how space alters cancer progression.

The newly detailed method allows for the creation of tiny replicas of the human brain so realistic that they rival “the complexity of the fetal brain’s neural network,” according to Muotri, who is also a professor in the UC San Diego School of Medicine’s Departments of Pediatrics and Cellular and Molecular Medicine. His brain replicas have already traveled to the International Space Station (ISS), where their activity was studied under conditions of microgravity.