Deep learning neural networks can be massive, demanding major computing power. In a test of the "lottery ticket hypothesis," MIT researchers have found leaner, more efficient subnetworks hidden within BERT models. The discovery could make natural language processing more accessible.
A machine learning model developed in Finland can help us treat cancer more effectively.
RoboGrammar is a new system that automates and optimizes robot design. The system, developed at MIT, creates arthropod-inspired robots for traversing a variety of terrains. It could spawn more inventive robot forms with enhanced functionality.
Researchers at the University of Maryland, Baltimore County (UMBC) have worked to create methods for improving the safety of technologically complex vehicles. Riadul Islam, assistant professor of computer science and electrical engineering, has worked with collaborators at UMBC and the University of Michigan-Dearborn to create a simple, easily adapted method for detecting the breaches in security.
MIT researchers have developed a way for deep learning neural networks to rapidly estimate confidence levels in their output. The advance could enhance safety and efficiency in AI-assisted decision making, with applications ranging from medical diagnosis to autonomous driving.
Scientists have designed a 2D material-based multi-stacked structure comprising tungsten disulfide (WS2) layer sandwiched between hexagonal boron nitride (hBN) layers that displays long-range interaction between successive WS2 layers with potential for reducing circuit design complexity and power consumption.
Researchers at MIT's Center for Bits and Atoms have created tiny building blocks that exhibit unique mechanical properties, such as the ability to produce a twisting motion when squeezed. These subunits could potentially be assembled by robots into a nearly limitless variety of objects with built-in functionality, including vehicles, large industrial parts, or specialized robots that can be repeatedly reassembled in different forms.
MCUNet is a new MIT system that brings machine learning to microcontrollers. The advance could enhance the function and security of devices connected to the Internet of Things (IoT).
Researchers from the University of Bristol's Quantum Engineering Technology Labs (QET Labs) and Université Côte d'Azur have made a new miniaturized light detector to measure quantum features of light in more detail than ever before. The device, made from two silicon chips working together, was used to measure the unique properties of "squeezed" quantum light at record high speeds.
University of Washington and Microsoft researchers have developed a DNA-based molecular tagging system.