{"slug": "catastrophic-compositional-generation-why-vanilla-diffusion-models-fail-to", "title": "Catastrophic Compositional Generation: Why Vanilla Diffusion Models Fail to Extrapolate", "summary": "Researchers argue that vanilla conditional diffusion models are fundamentally incapable of compositional generation when target distributions are out-of-distribution, as score estimation errors cause catastrophic performance degradation. Theory-guided generalization arguments and experiments on synthetic and realistic data support this claim, suggesting that inference-time techniques like Feynman-Kac correction cannot overcome the limitation.", "body_md": "arXiv:2606.23920v1 Announce Type: new\nAbstract: The task of compositional generation involves using a conditional generative model, trained only on a subset of the possible conditions, to produce samples from compositionally-defined target distributions such as a geometric combination of the source distributions. In this work, we argue that this task is often infeasible for vanilla conditional diffusion models: we conjecture that no inference-time technique can efficiently produce samples from the target distribution in certain well-motivated settings. This idea is supported by theory-guided generalization arguments and carefully-designed experiments on both synthetic and realistic data. In particular, while recent methods such as Feynman-Kac correction reduce inference-time approximation error, our results show that score estimation error has a more catastrophic effect on performance when the target distribution is out-of-distribution with respect to the sources, highlighting the need for a different approach to this task.", "url": "https://wpnews.pro/news/catastrophic-compositional-generation-why-vanilla-diffusion-models-fail-to", "canonical_source": "https://arxiv.org/abs/2606.23920", "published_at": "2026-06-24 04:00:00+00:00", "updated_at": "2026-06-24 04:29:34.412416+00:00", "lang": "en", "topics": ["machine-learning", "generative-ai", "ai-research", "neural-networks"], "entities": ["arXiv"], "alternates": {"html": "https://wpnews.pro/news/catastrophic-compositional-generation-why-vanilla-diffusion-models-fail-to", "markdown": "https://wpnews.pro/news/catastrophic-compositional-generation-why-vanilla-diffusion-models-fail-to.md", "text": "https://wpnews.pro/news/catastrophic-compositional-generation-why-vanilla-diffusion-models-fail-to.txt", "jsonld": "https://wpnews.pro/news/catastrophic-compositional-generation-why-vanilla-diffusion-models-fail-to.jsonld"}}