Iterative Feedback-Enhanced Prompting: A Green Algorithm for Reducing Household Food Waste

Large language models (LLMs) have demonstrated remarkable generative capabilities, but often struggle to encom-pass individual users' personal preferences. Existing method-ologies that work towards solving this exhibit drawbacks un-desirable for our use case. In this paper, we propose IFEP, an Iterative Feedback-Enhanced Prompting framework. IFEP is ever-improving and utilizes a preference selection algorithm with a weighted dynamic database to enhance LLM prompts for more user-personalized results. In our experiments, we observe improved user satisfaction over time in comparison to the commonly used GPT-3.5. We implement this framework in our improved household food tracking application, SmartFridge 2.0. With this app, we work to shape a sustainable world by mitigating the social issue of food waste.