School of Engineering Department of Computer Science and Engineering 80 Generative AI in Interactive Multimedia Applications Supervisor: BRAUD Tristan Camille / CSE Student: BOW Wing Yin / COMP Course: UROP 1000, Summer This project delves into the utilization of generative AI responses within the realm of games, exploring their applications and implications. Through this exploration, we strive to uncover the vast potential of generative AI in reshaping the landscape of gaming experiences, paving the way for new paradigms in interactive entertainment and fostering a deeper understanding of the evolving relationship between AI and human creativity. By delving into this realm, we aim to enrich our comprehension of how generative AI stands to transform interactions in the multimedia domain. This comprehensive report encapsulates our ongoing inquiry, illuminating the dynamic and innovative crossroads where artificial intelligence and human engagement converge. Generative AI in Interactive Multimedia Applications Supervisor: BRAUD Tristan Camille / CSE Student: FISILO William Arvin / ISD Course: UROP 3200, Fall The goal of this project is to explore the application of generative AI in innovative multimedia applications, focusing on virtual AI agents, world-building and gaming, storytelling, and musical interaction. The specific research objective is to develop NPC (non-player character) behaviour and visual representation based on user context input, creating a dynamic scheduler that can be modified by the user to suit their specific needs. Additionally, the project aims to study the interaction between objects and characters, visualize objects in the game environment, and convert the generated schedules into in-game objects. The report will begin with background research, then a simple prototype and finally, feedback from users. Generative AI in Interactive Multimedia Applications Supervisor: BRAUD Tristan Camille / CSE Student: HO Nga Kiu Kelly / CPEG Course: UROP 1100, Spring In this project, we investigated Generative artificial intelligence (GenAI) operation under a controlled environment. We set up a game scenario and introduce AutoGen agents as NPC to test out whether the agents could simulate a human conversation, while following the game flow, more specifically on identifying quest status and generating corresponding responses. The agents in this project could perform different tasks, including chatting with player, identify quest states, make decision on giving reward or not, to simulate a game flow. The experiment includes two versions of code, one makes use of AutoGen built-in functions, and another includes more self-defined functions and agents’ responses. Both give a positive result on following the game flow, while further development is still needed on system development and prompt structure.
RkJQdWJsaXNoZXIy NDk5Njg=