Data scientists, CISOs, entrepreneurs, and everyone else who wants a snapshot of the promise and potential of machine learning should check out the stirring talk delivered by Jeremy Howard at TEDxBrussels.
In his talk, Howard, who is a renowned Australian data scientist and entrepreneur, shared a brief overview of some of machine learning’s greatest hits including:
- The seminal moment in the 1962 when a computer used machine learning to beat a champion checkers player.
- The 2003 triumph of IBM’s Watson over two Jeopardy world champions.
- The growing reliance of global enterprises — including Google, Amazon Netflix, LinkedIn, and Facebook – on machine learning as a way to anticipate and influence customer behavior.
- The emergence of self-learning cars, including one that has already driven over 1 million accident-free miles on regular roads.
- The use of machine learning in a growing number of text-to-speech systems, as well as visual systems that are being used to automatically and rapidly recognize pictures, videos, and even physical locations.
- The growing use of the “deep learning” algorithm, which is inspired by how the human brain works and has already been used to discover pharmaceutical drugs in remarkably short time-frames, as well as identify new clinically relevant features that are helping doctors make better and faster prognosis.
According to Howard, “The better computers get at intellectual activities, the more they can build better computers to be better at intellectual capabilities, so this is going to be a kind of change that the world has actually never experienced before, so your previous understanding of what’s possible is different.”
Howard wraps up his talk by noting that machine learning is unquestionably a revolution of possibilities. He goes on emphasize that machine learning is “a case not where the human is being replaced by a computer, but where they’re working together. What we’re doing here is we’re replacing something that used to take a team of five or six people about seven years and replacing it with something that takes 15 minutes for one person acting alone.”
At Seculert, we agree with Howard that machine learning is not just altering how things are being done, but on a deeper level, it is changing what is fundamentally possible.
In cyber security, malware profiles are a critical input for log analysis. But because malware is evolving and new malware is appearing all the time, they are not enough to identify all threats. Seculert’s breach detection platform uses sophisticated machine learning algorithms to examine statistical features, classify the traffic, and determine whether it is similar to any of the known malware profiles.