Machine Learning (ML) algorithms require extremely large volumes of data to learn. This makes them susceptible to cyber threats. Consequently, there is a need to bolster cybersecurity in systems using ML techniques without hampering the performance. A new report published by The European Union Agency for Cybersecurity covers these cybersecurity aspects of machine learning. The report offers insights on preventing machine learning cyberattacks and deploying controls without hampering performance.
The Securing Machine Learning Algorithms report presents a taxonomy of ML techniques and core functionalities. The report maps the threats targeting ML techniques and the vulnerabilities of 40 commonly used ML algorithms. It provides a list of relevant security controls recommended to enhance cybersecurity in systems relying on ML techniques.
One of the challenges highlighted in the report is how to select the security controls to apply without jeopardizing the expected level of performance. The non-exhaustive machine learning algorithms taxonomy featured in the report supports the process of identifying which specific threats target ML algorithms, what are the associated vulnerabilities, and the security controls needed to address those vulnerabilities.
Click here to read the original article published by ENISA.
Please give your feedback on this article or share a similar story for publishing by clicking here.
