Science

How faster delivery of goods jeopardizes climate progress: study finds 20% lower efficiency of federal regulations

How faster delivery of goods jeopardizes ...

A new study shows that federal regulations on the energy efficiency of heavy trucks may be less effective than expected. Cheaper truck transport encourages the transition from rail to truck transport, increasing greenhouse gas emissions.

Detected bacteria that can destroy per- and polyfluoroalkyl substances, an important step towards more favorable purification of contaminated drinking water

Detected bacteria that can destroy per- and ...

A team of experts from the University of California Riverside has discovered bacteria from the genus Acetobacterium that can break the stable chemical bonds of PFAS, making a significant breakthrough in the treatment of contaminated drinking water sources.

Innovative gelatin batteries for wearables and medical implants: research from Cambridge University

Innovative gelatin batteries for wearables ...

Scientists at Cambridge University have developed stretchy gelatin batteries, inspired by electric eels, that could revolutionize wearable devices and medical implants. These batteries combine stretchability and conductivity in one material.

Water lentil: potential revolution in high-protein diet from simple freshwater plants

Water lentil: potential revolution in ...

Water lentil, a small aquatic plant rich in protein, has the potential to transform human nutrition. Scientists have developed methods to extract high-quality protein, which could significantly contribute to a sustainable and nutritious food future

A new approach to improve uncertainty assessment in machine learning models: a scalable method for applications in healthcare and other critical areas

A new approach to improve uncertainty ...

Mit researchers have developed an effective way to improve estimates of machine learning uncertainty, enabling more accurate and faster results in applications such as healthcare. This method helps users make informed decisions based on model reliability.

Developing a technique to assess the reliability of foundation models prior to application to specific tasks from MIT and MIT-IBM Watson AI Laboratory

Developing a technique to assess the ...

Researchers from mit and the MIT-IBM Watson AI Laboratory have developed a new technique to assess the reliability of foundation models before applying them to specific tasks, using an algorithm to assess model consistency. This solution can help reduce errors in safety-critical situations and enable better model selection without the need to test on actual data.