Machine learning can speed up response to cyber security threats and also reduce the risks, an e-book by Microsoft has found.

On average, attackers spend 200 days – that’s more than six months – sitting in networks, completely undetected. Machine learning can help increase detection of security threats to near real-time, the e-book says.

Governments can take advantage of a huge amount of threat intelligence they receive from across their own networks and that of private sector partners. Data from past attacks can be analysed to continuously improve your cyber defences using machine learning.

Using a combination of automated processes and human expertise, organisations can create an intelligence security graph that learns from itself and evolves, getting better at responding to cyber attacks every time.

If you would like to learn more about the use of machine learning in security from Microsoft, download the full e-book below.