.png){kind=icon}
.png){kind=icon}
Return
Evolutionary Content Generation via Multimodal LLM-Based Fitness Evaluation
.png){kind=icon}
PRE
AI
Y
N
Z
D
Y
张
Y
DOI：10.1109/tevc.2026.3657774.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Evolutionary algorithms (EAs) have gained prominence as a powerful optimization tool inspired by biological evolution, excelling in various complex domains. In the context of generative artificial intelligence (GAI), EAs have shown promise in generating diverse, high-quality solutions. However, traditional EAs heavily rely on human-designed fitness functions, which may often lack flexibility and comprehensiveness for different GAI scenarios. Recently, the emergence of large language models (LLMs) has opened new avenues for enhancing the evolutionary process. This article proposes a formal framework named evolutionary content generation framework with LLM-guided fitness evaluation (ECG-LFit), which utilizes an LLM (e.g., GPT-4-Turbo) for multimodal fitness evaluations. We validate the effectiveness of our framework using EAs (CMA-ES, MAP-Elites, and CMA-ME) in our case study on Super Mario game level generation. The results show that our framework improves the quality and playability of the generated levels. In addition, user studies indicate that participants prefer the levels generated by the ECG-LFit framework, particularly regarding aesthetics, challenge, and playability. Furthermore, to enhance the evaluation efficiency, we design a distilled model to simulate the scoring process of the LLM, enabling rapid and effective content evaluation in resource-constrained environments.
Keywords:
Distilled models
evolutionary algorithms (EAs)
fitness evaluation
generative artificial intelligence (GAI)
large language models (LLMs)
Journal
IF:
12
Papers:
1.8K
Citations:
2.4W
Organization
Cited Papers