Computers and Math from Science Daily Feed

Researchers find large language models process diverse types of data, like different languages, audio inputs, images, etc., similarly to how humans reason about complex problems. Like humans, LLMs integrate data inputs across modalities in a central hub that processes data in an input-type-agnostic fashion.

Scientists across Africa, America, Asia, Australia, and Europe outline for the first time how Artificial Intelligence (AI) can transform the landscape of infectious disease research and improve pandemic preparedness.

Physicists have developed a novel approach to maintain special quantum characteristics, even in 3D materials, with potential applications in optical systems and advanced computing.

A quantum 'miracle material' could support magnetic switching, a team of researchers has shown.

The efficiency of wireless charging systems is limited by power loss occurring due to frequency changes in the resonant circuits that enable power transfer. These necessary modulations reduce electromagnetic interference caused by resonant frequencies on other devices. However, conventional strategies for adapting to changing frequencies are inefficient, cost-prohibitive, and impractical. Now, scientists have designed a resonant tuning rectifier that provides a low-cost, efficient solution to stabilize power delivery in wireless power systems.

AI is reshaping workplaces, particularly in retail. Researchers explored how AI service quality impacts retail employees' innovation, job fit, and satisfaction. Findings show when employees perceive AI as reliable and empathetic, they are more likely to engage in innovative behavior. AI's adaptability also plays a crucial role in enhancing service quality. While reliability strongly supports innovation, transparency and responsiveness had less influence than expected. Empathy in AI systems was found to have a significant positive effect on employee innovation, creating a more engaging work environment. The study underscores AI's potential to drive service innovation in retail.

Ballbots are versatile robotic systems with the ability to move around in all directions. This makes it tricky to control their movement. In a recent study, a team has proposed a novel proportional integral derivative controller that, in combination with radial basis function neural network, robustly controls ballbot motion. This technology is expected to find applications in service robots, assistive robots, and delivery robots.

Using a novel surface-sensitive spectroscopy method, scientists explored atomic vibrations in crystalline material surfaces near interfaces. The findings illuminate quantum behaviors that play important roles computing and sensing technologies.

Are we putting our faith in technology that we don't fully understand? A new study comes at a time when AI systems are making decisions impacting our daily lives -- from banking and healthcare to crime detection. The study calls for an immediate shift in how AI models are designed and evaluated, emphasizing the need for transparency and trustworthiness in these powerful algorithms.

Researchers have demonstrated a new technique that uses light to tune the optical properties of quantum dots -- making the process faster, more energy-efficient and environmentally sustainable -- without compromising material quality.

In the world of soft robotics and wearable technology, sheet-based fluidic devices are revolutionizing how lightweight, flexible and multifunctional systems are designed. But with innovation comes challenges, particularly in understanding and controlling failure in these devices. A new study by mechanical engineers explores how programmed failure in heat-sealable, sheet-based systems can be used to protect devices, enable complex sequencing of actions and even streamline control mechanisms.

An innovative circuit design could enable miniature devices, such as microdrones and other microrobotics, to be powered for longer periods of time while staying lightweight and compact. Using miniaturized solid-state batteries, the circuit combines high energy density with an ultra lightweight design.

Research scientists in Switzerland have developed and tested a robust AI model that automatically segments major anatomic structures in MRI images, independent of sequence. In the study, the model outperformed other publicly available tools.

Using mathematical modeling, researchers have discovered that rate-induced tipping, which can happen if an environment changes too fast, can happen even in Daisyworld, a simple daisy-filled ecological model. If the planet heats up or cools down too quickly, all the daisies will go extinct, even if they would otherwise have been able to survive just fine under those conditions. This discovery mirrors similar observations found in other models and observed in real-life ecosystems.

Researchers have explored a 'quantum-inspired' technique to make the 'ones' and 'zeroes' for classical computer memory applications out of crystal defects, each the size of an individual atom. This turns milimeter-sized crystals into computer memory devices capable of storing terabytes of data. This interdisciplinary revolution in computer memory took its inspiration from the radiation dosimeters worn by hospital employees working with X-ray machines.

Three innovative design techniques substantially enhance wireless transmitter performance and can boost power efficiency and elevate data rates concurrently. This effectively aligns with the growing demand for speed and efficiency, accelerating the widespread deployment of wireless devices. This enables synergistic operation of wireless electronic devices and better quality of modern life.

A survey of adults found most had low trust in their health care system to use artificial intelligence responsibly or to make sure an AI tool would not harm them.

The Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) and NTT Research, Inc., a division of NTT, announced the publication of research showing an application of machine-learning directed optimization (ML-DO) that efficiently searches for high-performance design configurations in the context of biohybrid robots. Applying a machine learning approach, the researchers created mini biohybrid rays made of cardiomyocytes (heart muscle cells) and rubber with a wingspan of about 10 mm that are approximately two times more efficient at swimming than those recently developed under a conventional biomimetic approach.

A research team has developed two new autonomous navigation systems for cyborg insects to better navigate unknown, complex environments. The algorithms utilized only simple circuits that leveraged natural insect behaviors, like wall-following and climbing, to navigate challenging terrain, such as sandy, rock-strewn surfaces. For all difficulties of terrain tested, the cyborg insects were able to reach their target destination, demonstrating the potential of cyborg insects for surveillance, disaster-site exploration, and more.

Is it possible to tile a surface with a single shape in such a way that the pattern never repeats itself? In 2022, a mathematical solution to this 'Einstein problem' was discovered for the first time. Researchers have now also found a chemical solution: a molecule that arranges itself into complex, non-repeating patterns on a surface. The resulting aperiodic layer could even exhibit novel physical properties.