Authors:
Carlos Javier Hernández-Castro
1
;
Arturo Ribagorda
1
and
Julio Cesar Hernandez Castro
2
Affiliations:
1
Univ. Carlos III, Spain
;
2
Portsmouth University, United Kingdom
Keyword(s):
CAPTCHA, HIP, Common sense, AI, Logic reasoning, Semantic extraction, Data mining, Natural language modeling.
Related
Ontology
Subjects/Areas/Topics:
Access Control
;
Data Engineering
;
Databases and Data Security
;
Human Factors and Human Behaviour Recognition Techniques
;
Information and Systems Security
;
Information Assurance
;
Internet Technology
;
Network Security
;
Web Information Systems and Technologies
;
Wireless Network Security
Abstract:
CAPTCHAs or HIPs are tests able to tell humans and computers apart, remotely and over an untrustworthy channel. They rely on abilities that are though to be hard for algorithms, yet easy for humans. General logic reasoning, based on common sense knowledge, is one of the areas that are still considered hard for AI. On the other hand, logic reasoning targeting very specific areas has achieved success in AI. In this article, we list current Semantic and Logic CAPTCHAs and examine how strong they are. We also discuss wether this model is suited or not for automatic challenge generation and grading.