CLEAR: Cue Learning using Evolution for Accurate Recognition Applied to Sustainability Data Extraction
This addresses the need for domain experts in extracting sustainability data from building images, offering a significant improvement over manual methods.
The paper tackled the problem of improving LLM image recognition accuracy for specialized tasks by automatically generating and optimizing cues, achieving error rates up to two orders of magnitude lower than expert human recognition and human-authored prompts.
Large Language Model (LLM) image recognition is a powerful tool for extracting data from images, but accuracy depends on providing sufficient cues in the prompt - requiring a domain expert for specialized tasks. We introduce Cue Learning using Evolution for Accurate Recognition (CLEAR), which uses a combination of LLMs and evolutionary computation to generate and optimize cues such that recognition of specialized features in images is improved. It achieves this by auto-generating a novel domain-specific representation and then using it to optimize suitable textual cues with a genetic algorithm. We apply CLEAR to the real-world task of identifying sustainability data from interior and exterior images of buildings. We investigate the effects of using a variable-length representation compared to fixed-length and show how LLM consistency can be improved by refactoring from categorical to real-valued estimates. We show that CLEAR enables higher accuracy compared to expert human recognition and human-authored prompts in every task with error rates improved by up to two orders of magnitude and an ablation study evincing solution concision.