Explorations of an ai infused paradigm for education: tippae

Angel, John
Thumbnail Image
Other Contributors
Govindarajulu, Naveen, S
Nirenburg, Sergei
Hendler, Jim
Bello, Paul
Bringsjord, Selmer
Issue Date
Computer science
Terms of Use
This electronic version is a licensed copy owned by Rensselaer Polytechnic Institute (RPI), Troy, NY. Copyright of original work retained by author.
Full Citation
Both international and domestic data indicates that the US public education system is failing to teach its students mathematics. Not only this, but this data indicates that long-standing issues in the education system, such as performance gaps between socio-economic strata, have not only been growing worse with time in this domain but have been severely accelerated and exacerbated by Covid19. In this dissertation, I advance the creation of a paradigm of artificial agents designed to utilize several core pedagogical properties to tackle this education crisis. These advancements take the form of the creation and investigation of two novel formalizations that are paramount to the foundational properties of the paradigm as a whole. The first of these formalizations is the Artificial Agent Identity Teleportation Theory (AIT Theory for short) for migratable artificial intelligences which organizes the current state of the art on migratable agent identities into a unifying framework in computational logic and axiomizes the relationship between dynamic embodiment and presentation of identity. The second of these formalizations is the Logico-Mathematical Item Difficulty Theory (LID Theory for short) which captures a measure of the inherent difficulty of test items by examining the question's cognitive features through the medium of analysing the solution proof in computational logic. After introducing these formalizations, this work investigates the potential practical feasibility in utilizing them in real-time application environments through the engineering and analysis of prototype core functionalities of the paradigm being advanced. These core functionalities were built using logicist artificial intelligences capable of utilizing automated reasoning over the novel theories. The first of these core functionalities, the ability to take idealized perceptual information about an environment and recognize identity teleportation from one embodiment to another, can be executed in less than three seconds. The second core functionality, the ability to take in the information of fourth grade standardized testing math problems and estimate a numerical score of difficulty, executes in less than a second per problem across a wide range of subjects and difficulty levels. The results of these investigations reveal that real-time applications that exhibit correct behavior as defined by the novel theories are, indeed, practically feasible.
School of Science
Dept. of Computer Science
Rensselaer Polytechnic Institute, Troy, NY
Rensselaer Theses and Dissertations Online Collection
Restricted to current Rensselaer faculty, staff and students in accordance with the Rensselaer Standard license. Access inquiries may be directed to the Rensselaer Libraries.