Ph.D. Theses

Psychometric AI: A Test-based Approach Towards an Achievable Artificial Intelligence

By Bettina Schimanski
Advisor: Selmer Bringsjord
August 25, 2006

In an effort to build a system capable of human-level intelligence, what is the proper assessment for having achieved it, or preceding that, what exactly is the ultimate goal to strive toward? No universally accepted definition for human intelligence exists, let alone an undisputed account of artificial intelligence. Most current research in AI takes the form of stand-alone programs that perform one task well, but stumble or outright fail on other seemingly easy tasks.

The focus of this defense is twofold:

1) On the one hand, in a similar fashion to the Turing Test, and in line with Alan Newell's so-called "third paradigm" for addressing the study of the mind, we propose a standard to drive AI research and evaluate whether it has been successful. We are endeavoring to articulate and defend a new form of AI based, fundamentally, on the notion of well-defined tests. We call this type of Artificial Intelligence Psychometric AI.

2) In this venture, we have engineered the robot PERI so that it can successfully meet certain physical manipulation tests that are frequently issued as tests to humans to judge their facility with reasoning and motor control. We now turn our focus to the class of tests which include the challenge of predicting future actions from recorded, past behavior, such as the Picture Arrangement task from the current and renowned psychological test, the Wechsler Adult Intelligent Scale (WAIS). The challenge is to arrange snapshot images in an order that makes a plausible story or segment thereof. Why would solving such a task be significant? Among other reasons, it would lay the foundation for systems that are able to divine a plot that nefarious individuals are up to from only snapshots. A system competent in this ability would have great predictive power, a quality highly sought after in the security/defense world. Furthermore, we are making a contribution to the sub-field of AI that centers around stories, which includes story understanding, story generation, story completion, and story arrangement. Also, little research is currently focused on engineering an intelligent system capable of obtaining high-level information from diagrams, and of using such information in further problem solving. The project described herein changes that as we present our results on several Picture Arrangement problems.

Return to main PhD Theses page