Skip to main content

Machine Learning and maxing out at 77%

By March 9, 2019November 21st, 2019No Comments

Machine Learning (ML) is a sub-domain of Artificial Intelligence. It is part of an Artificial Intelligence (AI) solution. ML is not the answer but a piece of the solution. ML, based on studies at MIT AI labs, maxes out at a 77% success rate. This means a solution of Machine Learning fails at least 1 in 5 times.

When CCG compares Phen, a humanistic Strong AI system , to Machine Learning, this is the limitations of Machine Learning we are stating.

MIT AI labs have studied success rates of Machine Learning systems.

  • Unsupervised Machine Learning hits around 7% top end success.
  • Supervised Machine Learning gains success up to around 77%.

This means when utilizing Machine Learning as your security solution, 1 in 5 attempts go undetected. Machine Learning in security, just like security is an evolution, not a solution. There is no silver bullet.

Even all the Telsa driven hype behind their “auto-drive”, which still requires a human to maintain control, doesn’t get beyond this. Read more information in these two links

This is why CCG designed and developed CheckMate with Phen. Phen provides the dynamic abilities of human intellect at the speed of silicon. At the same time, CheckMate is a modular system that works with existing IT systems but provides more advanced security capabilities.

This allows CheckMate to adjust and expand as the bad guys develop new attack methods. Phen is using CheckMate and your IT systems to manage this security chess match.

CCG is providing insurance and protection against past, current and future attacks to your business. Phen is on the job so you can sleep at night.

Ray Canfield

Leave a Reply