New attack can make AI ‘see’ whatever you want
Peer-Reviewed Publication
Updates every hour. Last Updated: 23-Aug-2025 00:11 ET (23-Aug-2025 04:11 GMT/UTC)
Researchers have demonstrated a new way of attacking artificial intelligence computer vision systems, allowing them to control what the AI “sees.” The research shows that the new technique, called RisingAttacK, is effective at manipulating all of the most widely used AI computer vision systems.
MIT researchers developed a new system that enables a robot to use reflected Wi-Fi signals to identify the shape of a 3D object that is hidden from view, which could be especially useful in warehouse and factory settings.
GABA, or gamma-aminobutyric acid, is an amino acid functioning as the principal inhibitory neurotransmitter that can act on the brain to slow or stop the reception of certain signals to the brain, leading to a calmer and more relaxed state. Low GABA levels in the brain have been associated with neurological disorders and diseases like depression, Alzheimer's or epilepsy. Recently, there has been a push towards understanding more about the gut’s influence on mood, behavior and mental health, as well as what foods might fuel or hinder a healthy mind. Researchers set to work on determining whether brain GABA levels can be increased through dietary additions with the aim of modulating the gut bacteria present in an individual to bypass the blood-brain barrier, a barrier in which it is not proven yet GABA can pass through.
As global airspace regulations evolve and drone usage surges, accurately tracking multiple UAVs in dynamic swarm formations has become a pressing challenge for aerial safety, urban air mobility, and counter-drone operations. A research team from Beijing Institute of Technology has developed a novel visual tracking framework that significantly improves the identification and tracking of visually similar drones with nonlinear motion in air-to-air scenarios. Their work marks a major step toward scalable, intelligent swarm drone monitoring in real-world applications.
With the increasing focus on the pursuit-evasion game, the guidance law capturability analysis has been widely studied recently to theoretically assess the performance of different guidance laws and reveal the impact of the physical constraints on capture zones. In a recent study, the capture zones of the continuous and pulsed guidance laws in the pursuit-evasion game are analytically discussed to provide deep insights into the capturability distinction between the continuous guidance law and the pulsed guidance law.