Machine Learning vs Spam

We've all, at some stage, received these unwanted emails advertising products and services we most certainly have no interest in. Early attempts at blocking spam in the early 2000s produced a limited result. By creating antispam signatures – effectively manually created rules – the spam tide was only minimally stemmed with 50% of billions of spam mails being blocked while others still slipped through the net. The solution was found in machine learning, which analyses immense amounts of data and works out particular patterns. What resulted was the overall enhancement of defence and a catch of around 95% of spam.