Title :
PAAKL: Password Authentication Using Behavioral Metrics
Author :
Toptsis, Anestis A. ; Majonis, Joshua
Author_Institution :
Dept. of Comput. Sci. & Eng., York Univ., Toronto, ON, Canada
Abstract :
We present PAAKL, a method for password authentication. In addition to checking for a password´s correctness, PAAKL performs authentication by verifying that password using behavioral metrics extracted from the typing style of the rightful owner of a computer account. As such, PAAKL will deny access to users who have knowledge of a password but their typing style of that password is different from the typing style of the rightful owner. We implement and test our method with actual users. Our results indicate that the rightful owners of a password self-authenticate 92.5% of the time while the intruders - users that know the password of a rightful owner but are not rightful owners themselves, have 0% success in gaining access to a system.
Keywords :
authorisation; behavioural sciences; software agents; PAAKL; artificial k-line; behavioral metrics; computer account; password authentication; rightful owner; typing style; user access denied; Artificial intelligence; Artificial neural networks; Authentication; Companies; Delay; Pattern matching; Training; Artificial K-lines; Password authenticatiom;
Conference_Titel :
Computer Software and Applications Conference (COMPSAC), 2010 IEEE 34th Annual
Conference_Location :
Seoul
Print_ISBN :
978-1-4244-7512-4
Electronic_ISBN :
0730-3157
DOI :
10.1109/COMPSAC.2010.70