Date of Award

8-2026

Degree Type

Thesis

Degree Name

M.S.

Degree Program

Computer Science

Department

Computer Science

Major Professor

Shreya Banerjee

Second Advisor

Ben Samuel

Third Advisor

MD Meftahul Ferdaus

Abstract

Major advances in artificial intelligence have repeatedly promised general intelligence, yet each paradigm has encountered architectural limits. Despite rapid progress, the architectures, environments, and evaluation methods sufficient for artificial general intelligence (AGI) remain uncertain. This thesis proposes a neurocognitively inspired framework, defined at the computational and algorithmic levels of Marr’s analysis, to guide and assess AGI research. It reviews state-of-the-art vision-language models (VLMs) and world models, then introduces two hybrid architectures combining the semantic understanding of VLMs with the physical modeling capabilities of world models. These systems are evaluated in two-dimensional discrete and three-dimensional continuous environments, with future extensions to broader tasks. The thesis further argues that social simulations are essential AGI testbeds because they support open-ended environments, complex roles, multi-agent coordination, and emergent behavior under increasingly realistic social, cognitive, and physical constraints. Such simulations enable reciprocal human–AI learning, safety assessment, and applications in the real world.

Rights

The University of New Orleans and its agents retain the non-exclusive license to archive and make accessible this dissertation or thesis in whole or in part in all forms of media, now or hereafter known. The author retains all other ownership rights to the copyright of the thesis or dissertation.

Download

Available for download on Sunday, August 01, 2027

Share

COinS