When PR and reality collide: The truth about machine learning in cybersecurity

Discussion in 'other security issues & news' started by itman, Apr 20, 2017 at 8:27 AM.

  1. itman

    itman Registered Member

    Joined:
    Jun 22, 2010
    Posts:
    3,725
    Location:
    U.S.A.
    https://www.welivesecurity.com/2017/04/18/pr-reality-collide-truth-machine-learning-cybersecurity/
     
  2. itman

    itman Registered Member

    Joined:
    Jun 22, 2010
    Posts:
    3,725
    Location:
    U.S.A.
    http://www.darkreading.com/vulnerab...re-now-evades-machine-learning/d/d-id/1328506
     
  3. Peter2150

    Peter2150 Global Moderator

    Joined:
    Sep 20, 2003
    Posts:
    18,128
    No surprise to me.
     
  4. itman

    itman Registered Member

    Joined:
    Jun 22, 2010
    Posts:
    3,725
    Location:
    U.S.A.
  5. Umbra

    Umbra Registered Member

    Joined:
    Feb 10, 2011
    Posts:
    2,983
    Location:
    when i can counter-troll
    same here, there is no machine learning , it is just marketing crap , as if i add a new airbag mechanism in a car that can analyze the intensity of the impact and decide if it needs to be deployed by itself... i didn't add some kind of "intelligence" to the mechanism, just tweaked the threshold of the vibration sensor.... and this exactly what it is with ML , just more advanced heuristic algorithm, no "Ai" at all.
    You will have ML when you will have Skynet... :D
     
  6. XhenEd

    XhenEd Registered Member

    Joined:
    Mar 31, 2014
    Posts:
    479
    Location:
    Philippines
    There is machine learning. Machine learning, however, is only a subset of the whole Artificial Intelligence. ML, in itself, is not AI. From what I read, to form true AI, machine learning and other subsets/layers must be together, coordinating with each other. So, AI is the umbrella term, while ML, and other subsets, is only a part of it.

    The issue is whether ML is effective enough to combat malware, or that it is used properly by cybersecurity companies.

    Popular examples of machine learning are some chess programs.
     
  7. itman

    itman Registered Member

    Joined:
    Jun 22, 2010
    Posts:
    3,725
    Location:
    U.S.A.
    Important to prospective purchasers of this new technology is which methods of machine learning security software actually employ AI.

    Chess programs use AI algorithms whereas checkers programs do not not; they use pattern matching as noted in the link I posted in reply #4.
     
Loading...