Security News

Re: AI

Daily Dave - 1 April, 2016 - 09:20

Posted by Smoak, Christopher on Apr 01

I agree, and Anton, your blog post is very important for everyone to understand, especially consumers of these types of

My concern is that we're going to knee-jerk back into "ML is bad, period" because we have this notion of misuse (or
perhaps misguided use) within the industry. I know we all cringe when we hear the phrase "Big Data" these days, and
while there's certainly cause for such a reaction, it...

Re: AI

Daily Dave - 1 April, 2016 - 09:05

Posted by Sven Krasser on Apr 01

Your article has it right, we will need to ask ourselves that question in what cases black box detections are
desirable. In very general terms, in some cases there’s value in bringing instances to an analyst’s attention without
further reasoning (specifically if there is context available e.g. in the form of more forensic information that can be
accessed). In other cases interpretability is important (a very common problem in e.g....

Re: AI

Daily Dave - 1 April, 2016 - 08:51

Posted by Anton Chuvakin on Apr 01

.... but don't you guys [used generally] agree that ML and friends brings
up the challenge of "non-verifiability" to our domain [that I whine about
here, if you are curious].
Specifically, "because ML" argument is sometimes made not just by the
marketing droid [eh...I guess we can't use the word...

Re: AI

Daily Dave - 31 March, 2016 - 08:34

Posted by Allen DeRyke on Mar 31

For the time being I remain skeptical of ML "solutions" to intrusion
detection problems. In BJJ image processing its fairly simple for
humans to sanity check the results of ML. Humans are very good at
image processing which means we're pretty going to do a good job
spotting ML errors while working over the training data. Our ML
progress in the image processing space is a byproduct of our innate
biological adaptations for image...

Re: AI

Daily Dave - 31 March, 2016 - 08:22

Posted by Sven Krasser on Mar 31

Hey Chris,

We’re on the same page, and I think this is a healthy discussion to have :) Both an understanding and an open mind are
required to successfully use ML on complex and new problems.

Syndicate content