Date of Award


Document Type


Degree Name

Master of Science (MS)


College of Science and Mathematics


School of Computing

Thesis Sponsor/Dissertation Chair/Project Chair

Aparna Varde

Committee Member

Rui Li

Committee Member

Weitian Wang


This research explores the innovative integration of commonsense knowledge (CSK) within AI systems, with a particular focus on domestic robotics. At the heart of this study is the Robo- CSK-Organizer, a groundbreaking system that utilizes a classical knowledge base, namely ConceptNet, to enhance robotic decision-making through sophisticated object organization and classification. This system is contrasted with a ChatGPT-based organizer, examining their performance in terms of ambiguity resolution, consistency in object placement, adaptability to task classifications, and crucially, in explainability, a key aspect of XAI (Explainable AI). Through a combination of controlled experiments, quantitative and qualitative analysis, the study demonstrates that the Robo-CSK-Organizer not only classifies objects accurately, but also surpasses related systems such as the ChatGPT organizer in XAI. It dynamically applies CSK, thereby offering clearer and more understandable decision-making processes. These advancements are significant for the field of multipurpose robotics, suggesting enhanced user experience, increased trust, improved error detection and correction, as well as better comprehension of robotic actions by users. This thesis is distinguished by its comprehensive comparative analysis of CSK applications in AI, particularly highlighting the importance of XAI. It introduces an innovative framework for evaluating how AI systems resolve ambiguity and make decision-making processes transparent and interpretable. The study addresses the challenges of integrating CSK into robotics, focusing on ambiguity, consistency, task relevance, and explainability in object classification. It asserts that the Robo-CSK-Organizer, with its specialized deployment of a knowledge base (ConceptNet), not only achieves higher accuracy and consistency in object classification and placement but also provides superior explainability compared to a base case organizer that uses ChatGPT. The significance of this study lies in its contribution to AI research, offering empirical evidence on the strengths and limitations of CSK integration, especially in the context of XAI. It advances the conversation around AI transparency and demonstrates the practical application of CSK in multipurpose robotics. By illustrating how semantic task planning and CSK can be effectively operationalized in service robots with an emphasis on explainability, this research bridges an essential gap between theoretical concepts and their practical implementation in AI systems.

File Format


Available for download on Sunday, February 22, 2026