Toggle light / dark theme

The AI model that teaches itself to think through problems, no humans required

Artificial intelligence is getting smarter every day, but it still has its limits. One of the biggest challenges has been teaching advanced AI models to reason, which means solving problems step by step. But in a new paper published in the journal Nature, the team from DeepSeek AI, a Chinese artificial intelligence company, reports that they were able to teach their R1 model to reason on its own without human input.

When many of us try to solve a problem, we typically don’t get the answer straight away. We follow a methodical process that may involve gathering information and taking notes until we get to a solution. Traditionally, training AI models to reason has involved copying our approach. However, it is a long, drawn-out process where people show an AI model countless examples of how to work through a problem. It also means that AI is only as good as the examples it is given and can pick up on human biases.

Instead of showing the R1 model every step, researchers at DeepSeek AI used a technique called reinforcement learning. This trial-and-error approach, using rewards for , encouraged the model to reason for itself.

Magnetic tunnel junctions mimic synapse behavior for energy-efficient neuromorphic computing

The rapid development of artificial intelligence (AI) poses challenges to today’s computer technology. Conventional silicon processors are reaching their limits: they consume large amounts of energy, the storage and processing units are not interconnected and data transmission slows down complex applications.

As the size of AI models is constantly increasing and they are having to process huge amounts of data, the need for new computing architectures is rising. In addition to quantum computers, focus is shifting, in particular, to neuromorphic concepts. These systems are based on the way the works.

This is where the research of a team led by Dr. Tahereh Sadat Parvini and Prof. Dr. Markus Münzenberg from the University of Greifswald and colleagues from Portugal, Denmark and Germany began. They have found an innovative way to make computers of tomorrow significantly more energy-efficient. Their research centers around so-called magnetic tunnel junctions (MTJs), tiny components on the nanometer scale.

‘Quantum squeezing’ a nanoscale particle for the first time

Researchers Mitsuyoshi Kamba, Naoki Hara, and Kiyotaka Aikawa of the University of Tokyo have successfully demonstrated quantum squeezing of the motion of a nanoscale particle, a motion whose uncertainty is smaller than that of quantum mechanical fluctuations.

As enhancing the measurement precision of sensors is vital in many modern technologies, the achievement paves the way not only for basic research in fundamental physics but also for applications such as accurate autonomous driving and navigation without a GPS signal. The findings are published in the journal Science.

The physical world at the macroscale, from to planets, is governed by the laws of discovered by Newton in the 17th century. The physical world at the microscale, atoms and below, is governed by the laws of quantum mechanics, which lead to phenomena generally not observed at the macroscale.

Advanced AI links atomic structure to quantum tech

A research team led by Oak Ridge National Laboratory has developed a new method to uncover the atomic origins of unusual material behavior. This approach uses Bayesian deep learning, a form of artificial intelligence that combines probability theory and neural networks to analyze complex datasets with exceptional efficiency.

The technique reduces the amount of time needed for experiments. It helps researchers explore sample regions widely and rapidly converge on important features that exhibit interesting properties.

“This method makes it possible to study a material’s properties with much greater efficiency,” said ORNL’s Ganesh Narasimha. “Usually, we would need to scan a large region, and then several small regions, and perform spectroscopy, which is very time-consuming. Here, the AI algorithm takes control and does this process automatically and intelligently.”

Light-Powered AI Chips: The Photonic Revolution That’s About to Change Everything

Light-Powered AI Chips: The Photonic Revolution That’s About to Change Everything ## The future of artificial intelligence (AI) may be revolutionized by photonic AI chips that use light instead of electricity to process information, enabling faster, more efficient, and heat-free computing.

## Questions to inspire discussion.

Photonic AI Technology.

🔬 Q: What makes photonic AI chips more efficient than current AI chips? A: Photonic AI chips are 100x more energy efficient and produce virtually zero heat compared to electronic chips, as they use light instead of electrons for computation.

🌈 Q: How do photonic chips encode information differently? A: Photonic chips can encode information simultaneously in wavelength, amplitude, and phase by bouncing light off mirrors and optical devices, replacing traditional electronic processors.

Industry Developments.

Google DeepMind discovers new solutions to century-old problems in fluid dynamics

For centuries, mathematicians have developed complex equations to describe the fundamental physics involved in fluid dynamics. These laws govern everything from the swirling vortex of a hurricane to airflow lifting an airplane’s wing.

Experts can carefully craft scenarios that make theory go against practice, leading to situations which could never physically happen. These situations, such as when quantities like velocity or pressure become infinite, are called ‘singularities’ or ‘blow ups’. They help mathematicians identify fundamental limitations in the equations of fluid dynamics, and help improve our understanding of how the physical world functions.

In a new paper, we introduce an entirely new family of mathematical blow ups to some of the most complex equations that describe fluid motion. We’re publishing this work in collaboration with mathematicians and geophysicists from institutions including Brown University, New York University and Stanford University.

Spider-inspired magnetic soft robots could perform minimally invasive gastrointestinal tract procedures

The gastrointestinal (GI) tract is a collection of organs and structures inside the bodies of humans and other animals that is responsible for the digestion of food, the absorption of nutrients and the expulsion of waste. Its underlying parts include the mouth, esophagus, stomach, intestines, rectum and anus.

Over the past decades, the incidence of cancer in the GI tract and some other conditions affecting the digestive system has risen substantially. Existing approaches to diagnose and treat GI cancers rely on endoscopy, a medical procedure that entails the inspection of internal organs via a flexible tube with an embedded camera (i.e., endoscope), which is inserted into the body through the anus, mouth or a small incision.

In addition to being highly uncomfortable for patients, endoscopy often fails to reach regions that are deep into the GI tract or are difficult to access due to the body’s natural configuration. Some have thus been trying to devise alternative systems that could inspect parts of the digestive system more effectively, while causing patients minimal discomfort.

/* */