Generative Models in Decision Making: A Survey

2025-02-24

AI Summary

Existing research on generative models for decision-making lacks a unified conceptual framework and systematic taxonomy, hindering comparative analysis and practical deployment. Method: We propose the first comprehensive taxonomy unifying seven major generative model families—Energy-Based Models (EBMs), Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), Normalizing Flows, Diffusion Models, Generative Flow Networks (GFlowNets), and Autoregressive Models. We introduce a functional “Controller–Modeler–Optimizer” framework for generative decision-making, instantiated across five real-world domains including autonomous driving and medical diagnosis. We identify and formalize three key evolutionary trajectories: high-performance algorithms, large-scale generalizable models, and self-evolving, adaptive models. Contribution/Results: This work establishes the first complete classification system and standardized evaluation benchmark for generative decision-making. It clarifies methodological strengths, fundamental bottlenecks, and viable pathways for breakthroughs—providing both theoretical foundations and practical guidelines for trustworthy AI-driven decision systems.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

In recent years, the exceptional performance of generative models in generative tasks has sparked significant interest in their integration into decision-making processes. Due to their ability to handle complex data distributions and their strong model capacity, generative models can be effectively incorporated into decision-making systems by generating trajectories that guide agents toward high-reward state-action regions or intermediate sub-goals. This paper presents a comprehensive review of the application of generative models in decision-making tasks. We classify seven fundamental types of generative models: energy-based models, generative adversarial networks, variational autoencoders, normalizing flows, diffusion models, generative flow networks, and autoregressive models. Regarding their applications, we categorize their functions into three main roles: controllers, modelers and optimizers, and discuss how each role contributes to decision-making. Furthermore, we examine the deployment of these models across five critical real-world decision-making scenarios. Finally, we summarize the strengths and limitations of current approaches and propose three key directions for advancing next-generation generative directive models: high-performance algorithms, large-scale generalized decision-making models, and self-evolving and adaptive models.

Problem

• Integrate generative models into decision-making
• Classify seven types of generative models
• Examine models in real-world decision scenarios

Innovation

• Generative models classify into seven types
• Controllers, modelers, optimizers enhance decision-making
• High-performance algorithms advance generative directive models