2020-05-31 智邦网
编译 致远
据mil-embedded网站2020年5月30日报道,近日,城堡防御(Citadel Defence)公司发布采用深度神经网络技术软件“Deepfake”,帮助美国及其盟友对抗敌人多变的电子战欺骗战术。
该“泰坦–反无人机”(Titan C-UAS )系统使用生成对抗网络和专有的图像生成算法,通过深度学习技术设计出具备判别能力的分类模型,对检测到的信号是真正的无人机信号还是敌人模拟生成的欺骗性信号进行预判。
过去三个月中,城堡防御公司已对500多名军事和执法人员组织业务培训,主要学习新型信息对抗威胁战术和泰坦系统技术。同时,已有100多个泰坦系统正在更新软件版本。
Deepfake AI software developed for Titan C-UAS
NATIONAL CITY, Calif. Citadel Defense, full-spectrum counter-drone solutions company, released new software incorporating deepfake neural networks to protect against adversarial spoofing tactics.
This capability could help United States and allied forces combat growing enemy tactics that attempt to confuse existing security intelligence equipment as electronic warfare escalates.
By using Generative Adversarial Networks in the Titan C-UAS solution and proprietary image generation algorithms, Citadel has developed discrimination classification models intended to help determine whether the signal detected is a real drone or a generated signal by the adversary trying to trick existing signal intelligence equipment.
In the last three months, Citadel has trained over 500 military and law enforcement operators on emerging adversarial threat tactics and Titan technology, according to the company. The company also claims that over 100 Titan systems are being updated with the company’s latest software release.
相关新闻
美国会提1000亿美元议案 强化人工智能、量子计算、机器人技术研发(英文)
SparkCognition设立专业公司 开发新一代军用人工智能系统(英文)