Using Social Networking to Fight Bad Actors

Check out my cyber security presentation from GFIRST 2012...

GFIRST was a 1500-person conference hosted by the Department of Homeland Security. I conducted this presentation in round-robin fashion with my two partners in crime, Dave Roberts and Jonathan Quigg.



Comments

David Keener By Dave Aronson on Friday, September 20, 2013 at 07:17 PM EST

Hi David!

I just read your slides, and the concept is very reminiscent of some things I have suggested many times to LinkedIn. Groups are so flooded with spammy recruiters and other kinds of jerks, that they sorely need some kind of reputation system, along with group management tools to automagically eject or moderate people with a very low reputation (as opposed to not having established any rep yet).

I've given some thought to how the usual suspects could game the system, with tons of sockpuppet accounts Like-ing each others' content, and had the idea of lowering the impact of a like or dislike from someone with a bad rep, and sockpuppet detection by comparing their patterns of likes and dislikes to those of others. But that just made my head hurt. ;-)

-Dave




David Keener By dkeener on Sunday, September 22, 2013 at 09:04 PM EST

Hi Dave,

First, thanks for the excellent feedback on the slides. I'm glad you enjoyed them.

On the whole "gaming the system" problem, I've been there and done that. I implemented a video contest for the Department of State one time, with peer voting.

There were clear patterns of voting fraud that you could detect. You could therefore classify people as "responsible voters" or "bad voters", and discount their respective votes accordingly. We also had to deal with accounts like "voter01@yahoo.com", "voter02@yahoo.com", etc. It was terribly amusing.

            -- Dave --




David Keener By dkeener on Tuesday, January 21, 2014 at 08:19 PM EST

Just an FYI. This presentation has had over 1100 views on Slideshare.net. I'm particularly pleased that so many people have found it useful.


Leave a Comment

Comments are moderated and will not appear on the site until reviewed.

(not displayed)