AI Turns the Tables: Designing Ceramic-Metal Joints Backwards from Target Strength
Researchers at IIT Roorkee and ISRO have developed an AI-driven inverse design method that predicts which ceramic, braze alloy, and ...
Researchers at IIT Roorkee and ISRO have developed an AI-driven inverse design method that predicts which ceramic, braze alloy, and ...
A new neural network called CDGaitFusion fuses shared motion patterns with individual-specific features to recognize people by their gait despite ...
Researchers have developed a real-time AI framework and a new field-captured dataset, VerBeet, that localize sugar beets and detect breakage ...
Researchers have developed a single-arm robotic sewing system that combines computer vision, neural-network grasp estimation, and force control to handle ...
Researchers at EPFL have developed MAGNETS, an inherently interpretable neural network that predicts continuous targets from time series by learning ...
NanoTS, a new deep learning tool published in Nature Methods, enables accurate SNP detection and genotype calling from nanopore long-read ...
Researchers in Morocco have developed IACAN, a deep learning architecture that dynamically balances convolutional and attention branches using KL divergence ...
Researchers have developed a hypernetwork-based framework that lets continually learning AI systems erase specific tasks without access to the original ...
Researchers in China have developed a deep learning model that analyzes complete, variable-length opcode sequences from blockchain transactions, detecting seven ...
Researchers highlight reconfigurable transistor hardware that filters task-relevant visual inputs before computation, cutting wasted activations in neuromorphic AI systems.
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine