This thesis explores the integration of Artificial Intelligence as a Service (AIaaS) into existing systems, focusing on handling challenges related to unclear data processing and complex integration. The study examined existing research to understand current integration practices and ensure alignment with established standards. Based on this research, an integration architecture was designed and emphasizes two key factors: ensuring the system works as expected (functional suitability) and ensuring different parts of the system can communicate smoothly (interoperability). The integration architecture was designed to simplify communication between different parts of the system, making sure they all work together effectively. It also helps reduce the complications that often come with integration. This mutual reinforcement between functional suitability and interoperability implies coherent outcomes and establishes an environment that fosters smooth communication among system components. The practical implications of this research are exemplified through the implementation of the proposed architecture within the Gokind platform, resulting in positive outcomes. The transition from manual receipt verification to automated receipt recognition using the Google Vision Application Programming Interface (API) showcases accelerated processing times, scalability, and efficient resource allocation. Despite achieving an impressive 90% accuracy rate, the study identifies areas for potential improvement, advocating for ongoing refinement. While the study successfully navigates the challenges related to Artificial Intelligence as a Service (AIaas) integration, it acknowledges certain limitations, such as the potential for exploring varied AIaas providers and environments and the essential consideration of security aspects. Moreover, future research avenues are suggested, including variance analysis across AIaas classes, comparative studies among providers, fortified security measures, and comprehensive exploration of architectural attributes’ impact.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:mdh-64573 |
| Date | January 2023 |
| Creators | Musabimana Boneza, Benedicte |
| Publisher | Mälardalens universitet, Akademin för innovation, design och teknik |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0023 seconds