Guided Flows for Generative Modeling and Decision Making

About

Classifier-free guidance is a key component for enhancing the performance of conditional generative models across diverse tasks. While it has previously demonstrated remarkable improvements for the sample quality, it has only been exclusively employed for diffusion models. In this paper, we integrate classifier-free guidance into Flow Matching (FM) models, an alternative simulation-free approach that trains Continuous Normalizing Flows (CNFs) based on regressing vector fields. We explore the usage of \emph{Guided Flows} for a variety of downstream applications. We show that Guided Flows significantly improves the sample quality in conditional image generation and zero-shot text-to-speech synthesis, boasting state-of-the-art performance. Notably, we are the first to apply flow models for plan generation in the offline reinforcement learning setting, showcasing a 10x speedup in computation compared to diffusion models while maintaining comparable performance.

Qinqing Zheng, Matt Le, Neta Shaul, Yaron Lipman, Aditya Grover, Ricky T. Q. Chen · 2023

Related benchmarks

Task	Dataset	Result
Simulation-Based Inference	SBIBM Gaussian Linear	C2ST0.69	19
Simulation-Based Inference	Gaussian Linear	Computation Time (s)0.01	8
Simulation-Based Inference	Gaussian Mixture	Computation Time (s)0.01	8
Simulation-Based Inference	Bernoulli GLM	Computation Time (s)0.01	8
Simulation-Based Inference	Two Moons	Computation Time (s)0.01	8
Simulation-Based Inference	SLCP	Inference Time (s)0.01	8
Posterior Sampling	Gaussian Mixture SBI benchmark	C2ST85	7
Posterior Sampling	Bernoulli GLM SBI	C2ST88	7
Posterior Sampling	SLCP SBI benchmark	C2ST91	7
Posterior Sampling	Two Moons SBI benchmark	C2ST78	6