A group of scientists from cyber security division of IBM have generated an AI-supported biometric malware dubbed as DeepLocker. This malware can work with pictures harvested from social media, claimed the media.
“DeepLocker is a new type of highly targeted and highly evasive malware that basically is different from any malware that is present these days,” claimed Dr. Marc Ph. Stoecklin, Principal Research Scientist at IBM Research for Cognitive Cyber Security Intelligence, to the media in an interview. “In the future, things are going to be AI vs. AI,” Stoecklin claimed further.
The media also claimed that Trustwave (the cyber security management firm) has designed an open source tool for white hackers. This tool harvests pictures from LinkedIn and then searches other social media networks for matches. Linking a profile on LinkedIn (which fundamentally contains professional and employment data) with social media profiles on platforms such as Facebook can be helpful for rolling out targeted phishing attacks.
Speaking of IBM, the company earlier revealed that its next-gen power systems servers integrated POWER9, its newly-developed processor. Built particularly for compute-concentrated AI workloads, the new systems of POWER9 are able to improve the training times of deep learning structures by almost 4x permitting ventures to build more precise and faster AI applications.
“The latest Power Systems Servers of IBM with POWER9 processor will be effectual for deep learning and AI jobs. The latest processor gives on unparalleled cognitive abilities and can assist Indian enterprises all over the verticals to up-scale and transform on their machine learning and AI voyage,” claimed Director at Systems for India and South Asia, Viswanath Ramaswamy, to the media in an interview.
The system was developed to boost demonstrable performance enhancements all over popular AI structures such as TensorFlow, Chainer, and Caffe, as well as hurried databases similar to Kinetica. Consequently, data researchers can develop apps faster, tuning from real-time fraud detection, deep learning perceptions in scientific study, and credit risk analysis.