Diffusion Models with Double Guidance: Generate with aggregated datasets

TL;DR

This paper introduces a diffusion model with double guidance that enables precise conditional generation across aggregated datasets with inconsistent attributes, without needing joint annotations.

Contribution

It proposes a novel diffusion approach that maintains control over multiple conditions despite missing joint attribute annotations, improving generative controllability.

Findings

Outperforms baselines in molecular and image generation tasks.

Achieves better alignment with target distributions.

Demonstrates effective control with incomplete attribute data.

Abstract

Creating large-scale datasets for training high-performance generative models is often prohibitively expensive, especially when associated attributes or annotations must be provided. As a result, merging existing datasets has become a common strategy. However, the sets of attributes across datasets are often inconsistent, and their naive concatenation typically leads to block-wise missing conditions. This presents a significant challenge for conditional generative modeling when the multiple attributes are used jointly as conditions, thereby limiting the model's controllability and applicability. To address this issue, we propose a novel generative approach, Diffusion Model with Double Guidance, which enables precise conditional generation even when no training samples contain all conditions simultaneously. Our method maintains rigorous control over multiple conditions without requiring joint annotations. We demonstrate its effectiveness in molecular and image generation tasks, where it outperforms existing baselines both in alignment with target conditional distributions and in controllability under missing condition settings.