PRISM: Diversifying Dataset Distillation by Decoupling Architectural Priors

Brian Bernhard Moser, Shalini Sarode, Federico Raue, Stanislav Frolov, Krzysztof Adamkiewicz, Arundhati Shanbhag, Joachim Folz, Tobias Christian Nauen, Andreas Dengel

Keywords: Dataset Distillation, PRISM, Architectural Priors, Intra-class Diversity

TL;DR: PRISM makes distilled datasets more diverse and useful by using different neural networks for the classification task and for regularization, breaking the bottleneck of relying on a single model's limited "view" of the world.

Abstract: Dataset distillation (DD) promises compact yet faithful synthetic data, but existing approaches often inherit the inductive bias of a single teacher model. As dataset size increases, this bias drives generation toward overly smooth, homogeneous samples, reducing intra-class diversity and limiting generalization. We present PRISM (PRIors from diverse Source Models), a framework that disentangles architectural priors during synthesis. PRISM decouples the logit-matching and regularization objectives, supervising them with different teacher architectures: a primary model for logits and a stochastic subset for batch-normalization (BN) alignment. On ImageNet-1K, PRISM consistently outperforms single-teacher methods (e.g., SRe2L) and recent multi-teacher variants (e.g., G-VBSM) at low- and mid-IPC regimes. The generated data also show significantly richer intra-class diversity, as reflected by a notable drop in cosine similarity between features. We further analyze teacher selection strategies (pre- vs. intra-distillation) and introduce a scalable cross-class batch formation scheme for fast parallel synthesis. Code will be released after the review period.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Submission Number: 1816