aiqbench analysis
This research introduces Label-Augmented Conditional Diffusion (LACD), a method that leverages unlabeled data by assigning them trivial labels to improve sample efficiency in conditional diffusion models. Theoretically, the approach guarantees faster convergence than purely supervised estimators when ample unlabeled samples are available, with rigorous proofs of identifiability and statistical performance across image, tabular, and synthetic benchmarks. For practitioners facing high labeling costs, this technique offers a pathway to maximize generative model utility without sacrificing distributional accuracy.
Semi-Supervised Conditional Diffusion via Label Augmentation
arXiv:2607.16685v1 Announce Type: cross Abstract: Conditional diffusion models have become a powerful and flexible framework for learning complex conditional distributions from labeled data. In practice, however, acquiring high-quality labels is costly and time-consuming, leaving large volumes of unlabeled data unused. To address this, we introduce label-augmented conditional diffusion (LACD), a simple and effective approach that incorporates unlabeled examples by assigning them a designated trivial label and performing joint denoising score matching over the augmented dataset. We provide sufficient conditions guaranteeing population-level identifiability of the target conditional distribution under this scheme. Moreover, we establish rigorous statistical guarantees: when sufficiently many unlabeled samples are available, the sampling distribution produced by LACD converges strictly faster than the purely supervised estimator in total variation distance, and at least as fast in Wasserstein-1 distance. Extensive experiments on synthetic, image, and tabular benchmarks corroborate our theory and show substantial gains in sample efficiency and generative performance compared with the purely supervised estimator.
Abstract:Conditional diffusion models have become a powerful and flexible framework for learning complex conditional distributions from labeled data. In practice, however, acquiring high-quality labels is costly and time-consuming, leaving large volumes of unlabeled data unused. To address this, we introduce label-augmented conditional diffusion (LACD), a simple and effective approach that incorporates unlabeled examples by assigning them a designated trivial label and performing joint denoising score matching over the augmented dataset. We provide sufficient conditions guaranteeing population-level identifiability of the target conditional distribution under this scheme. Moreover, we establish rigorous statistical guarantees: when sufficiently many unlabeled samples are available, the sampling distribution produced by LACD converges strictly faster than the purely supervised estimator in total variation distance, and at least as fast in Wasserstein-1 distance. Extensive experiments on synthetic, image, and tabular benchmarks corroborate our theory and show substantial gains in sample efficiency and generative performance compared with the purely supervised estimator.
Machine Learning (stat.ML); Machine Learning (cs.LG)
arXiv:2607.16685 [stat.ML]
(or arXiv:2607.16685v1 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2607.16685 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 248 words
Related coverage
Conditional Evaluation of Language Models with Cheap Auxiliary Signals
arXiv · Aug 19, 2026
GMoT: Gated Motion-Aware Tokenization for Fine-Grained Micro-Gesture Video Reasoning with Multimodal LLMs
arXiv · Jul 21, 2026
Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk
Latent Space · Oct 8, 2026
Semi-Supervised Conditional Generative Learning through Stochastic Interpolation and Sufficient Representations
arXiv · Jul 21, 2026