news

Researchers have discovered key 'units' in large AI models that seem to be important for language, mirroring the brain's language system. When these specific units were turned off, the models got much worse at language tasks. Large Language Models (LLMs) are not just good at understanding and using language, they can also reason or think logically, solve problems and some can even predict the thoughts, beliefs or emotions of people they interact with.

In just 15 minutes, training with a simulator enabled professional field hockey players to significantly improve their perceptive skills in front of the opposing net. This new approach, developed by researchers at the University of Fribourg's Perception and Control Laboratory, enables players to better and more quickly identify the area of the goal least covered by the goalkeeper.

To the point The researchers used the video game Minecraft to study social learning processes in a dynamic, realistic environment. The study shows that adaptability-i.e., the flexible shift between individual and social learning-is crucial for success. Using new computer-based methods to track gaze data and model decision-making, the researchers were able to precisely describe and predict individual and social learning strategies.

Researchers have developed a method that makes AI responses increasingly reliable. Their algorithm specifically selects data relevant to the question. In addition, even AI models up to 40 times smaller achieve the same output performance as the best large AI models. ChatGPT and alike often amaze us with the accuracy of their answers, but unfortunately, they also repeatedly give us cause for doubt.

Researchers have developed the first AI model of language in the brain that captures both how neurons are arranged and how they function. Our brain is a very organized place. Neurons - the nerve cells responsible for transmitting electrical and chemical signals throughout the body - are organized on tissue, tending to cluster together in groups according to how they function.

Wearable technology aided by artificial intelligence can warn and prevent falls for patients recovering from stroke and spinal injuries. Simon Fraser researchers are using cutting edge technology, in combination with artificial intelligence and machine learning, to support safer rehabilitation for patients.

How can artificial intelligence enhance our understanding of the visual system in the brain? An international research team (MICrONS), with the participation of the University of Göttingen, has developed new AI models to decode the complex processing of visual stimuli in the brain. The researchers investigated how the shape, connectivity and activity of nerve cells in the mouse brain are related.

Researchers at Durham and Northumbria Universities have helped develop an app to better equip frontline practitioners to identify hate incidents and ensure victims get the right support. The Hate ID app will help practitioners in the UK Northumbria Police and Crime Commissioner region signpost victims of hate incidents to the most appropriate support service.

Researchers from Vrije Universiteit Brussel (VUB) and Universitair Ziekenhuis Brussel (UZ Brussel) have developed an innovative digital application that enables fast and efficient screening of cognitive and motor impairments in multiple sclerosis (MS) patients. The app provides a simple and accessible way to track cognitive decline and motor limitations, marking an important step in the digitalization of neurological care.

Researchers have developed a handheld device that could potentially replace stethoscopes as a tool for detecting certain types of heart disease. This device could become an affordable and scalable solution for heart health screening, especially in areas with limited medical resources Anurag Agarwal The researchers, from the University of Cambridge, developed a device that makes it easy for people with or without medical training to record heart sounds accurately.

Publication by the University of Bonn warns of misunderstandings in handling predictive algorithms Researchers from chemistry, biology, and medicine are increasingly turning to AI models to develop new hypotheses. However, it is often unclear on which basis the algorithms come to their conclusions and to what extent they can be generalized.

At the LCSB, researchers model the complexity of the human brain in a Petri dish. Their work is highlighted in the latest episode of Primer , a new documentary series created by Bloomberg Originals that analyzes the complex science and technology shaping the future. Broadcast on March 27, the episode is entitled "Can Living Human Brain Cells Power AI?".

New miniature laboratories are ensuring that artificial intelligence (AI) doesn't make mistakes. They provide a controlled test environment where algorithms and AI models can be checked before being put to work under real-life conditions. The aim is for AI to work reliably. Anyone who develops an AI solution sometimes goes on a journey into the unknown.

For the first time, researchers at the University of Waterloo's Institute for Quantum Computing (IQC) and the University of Innsbruck in Austria have performed a quantum simulation of a two-dimensional particle physics theory on a qudit quantum computer, bringing us closer to understanding nature at its most fundamental level.

Research into elementary particles and their interactions is of central importance for our understanding of the universe. A team from the Universities of Innsbruck and Waterloo shows how a new type of quantum computer opens a door into the world of particle physics. The standard model of particle physics provides the best description to date of the forces and particles that make up our world.

Every five minutes, somewhere in the world, someone creates a new deepfake. As this AI-powered threat grows, researchers are developing equally powerful tools to fight back. In 2024, a deepfake attempt occurred every five minutes, according to a report by the Entrust Cybersecurity Institute. The report shows a worrying trend.

First developed in 2018 by Professor Christophe Ley , this innovative approach to machine learning is making waves far beyond the lab. By blending statistical insights with machine learning models, Statistically Enhanced Learning unlocks new predictive power. Let's check in on how this brainchild is growing up.

A new AI weather prediction system, developed by researchers from the University of Cambridge, can deliver accurate forecasts tens of times faster and using thousands of times less computing power than current AI and physics-based forecasting systems. The system, Aardvark Weather, has been supported by the Alan Turing Institute, Microsoft Research and the European Centre for Medium Range Weather Forecasts.

The Euclid mission delivers its first batch of data-valuable information to help unravel, among other things, the mystery of dark matter. The European Space Agency's (ESA) Euclid mission has unveiled its first public data, offering unprecedented insights into the structure and evolution of the Universe.

When a person's hearing and vision are uncompromised and function at a relatively high level, the human brain is able to take in various sights and sounds from any environment and seamlessly allow said person to perceive what's happening around them. But how does it work? Spoiler alert: There's more than meets the eye.