Defense Advanced Research Projects AgencyTagged Content List


Relating to digital systems and information

Showing 110 results for Cyber RSS
The Memex program seeks to develop the next generation of search technologies and revolutionize the discovery, organization and presentation of search results.
The February 2011 Federal Cloud Computing Strategy released by the U.S. Chief Information Officer reinforces the United States Government’s plans to move information technology away from traditional workstations and toward cloud computing environments. Where compelling incentives to do this exist, security implications of concentrating sensitive data and computation into computing clouds have yet to be fully addressed. The perimeter defense focus of traditional security solutions is not sufficient to secure existing enclaves. It could be further marginalized in cloud environments where there is a huge concentration of homogeneous hosts on high-speed networks without internal checks, and with implicit trust among hosts within those limited perimeter defenses.
Cyberspace is now recognized as a critical domain of operations by the U.S. military and its protection is a national security issue.
Cloud computing provides computing capabilities as a service rather than a product. Advantages to this include reduced costs and maintenance, and increased flexibility, availability and scalability. Cloud computing, however, also presents some potentially significant security issues. In particular, vulnerabilities could include compromise of data security and loss of key information. Any computer or web-friendly device connected to the Internet could gain unauthorized access to pools of computing power, applications, or files – compromising information security in cloud-computing environments.
Researchers have demonstrated effective attacks on machine learning (ML) algorithms. These attacks can cause high-confidence misclassifications of input data, even if the attacker lacks detailed knowledge of the ML classifier algorithm and/or training data. Developing effective defenses against such attacks is essential if ML is to be used for defense, security, or health and safety applications.