I didn't see any distinction between types of "deception" in the paper which means it's classifying things like social anxiety and generally self-conscious beha

I didn't see any distinction between types of "deception" in the paper which means it's classifying things like social anxiety and generally self-conscious behavior as deceptive. While strictly true it makes the algorithm useless for detecting security threats at this stage.