OneFlowSBI is a masked flow matching framework that trains a single network on the joint distribution of parameters and observations, using a ternary mask to indicate observed, generated, or absent coordinates. This mask-based approach allows any conditional or marginal inference to be performed by simply changing the mask, avoiding the quadratic attention cost of transformer-based generalists. Experiments on high‑dimensional image denoising and shallow water inference show that OneFlowSBI achieves lower posterior error, faster sampling, and can handle missing data without retraining, while matching or exceeding neural posterior estimation on most SBIBM tasks.
arXiv:2605.16818v2 Announce Type: replace-cross Abstract: Learning physical fields directly from incomplete observations is challenging because authentic occlusions are structured, sample-dependent,...
The paper introduces Transformer-Within-Transformer (TWT), a post‑hoc technique that merges contiguous redundant layers in Vision Transformers into a single surrogate layer. By doing so, TWT cuts both parameter count and inference compute while maintaining competitive performance on natural image tasks with only half the depth. In histopathology applications, TWT not only matches but sometimes surpasses the baseline model’s performance.
arXiv:2509. 23385v5 Announce Type: replace-cross Abstract: Simulation-based inference (SBI) is transforming experimental sciences by enabling parameter estimation in complex non-linear models from simulated data.
Conditional Mean Flows introduce a new approach for simulation-based inference that replaces the costly numerical integration of traditional flow matching with few-step sampling and tractable density evaluation. By identifying conditional flow maps through a mean flow objective and using independent masking, the method allows direct estimation of marginal likelihoods. Experiments demonstrate that single-step conditional mean flow density evaluation achieves error comparable to 200-step flow matching while accelerating downstream MCMC sampling by roughly 200‑fold.
arXiv:2610. 09281v1 Announce Type: new Abstract: In Bayesian inverse problems, posterior sampling requires generating samples that are consistent with given observations while capturing the range of plausible solutions.
The paper investigates how one-step generative models, which compress multi-step diffusion processes, handle denoising across a single forward pass. It introduces the concept of "depth as time," showing that intermediate layers of a model can be decoded to reveal a denoising trajectory similar to multi-step diffusion. The study finds that models trained with a time-indexed transport task, such as MeanFlow, exhibit this depthwise denoising and can be highly compressed, whereas models without such a task do not show the same behavior.
arXiv:2605. 00941v4 Announce Type: replace Abstract: Flow matching has become a leading framework for generative modeling, but quantifying the uncertainty of its samples remains an open problem.
arXiv:2602.05391v3 Announce Type: replace Abstract: Dataset distillation seeks to synthesize a compact surrogate dataset that enables performance comparable to training on the original dataset for do...
arXiv:2609.12953v1 Announce Type: new Abstract: Flow matching approaches to imaging inverse problems commonly incorporate measurements in two ways. Conditioning-based approaches supply measurement-de...
arXiv:2512.11695v3 Announce Type: replace-cross Abstract: Particle Image Velocimetry (PIV) is among the central modalities for measuring flow fields across laboratory, industrial and environmental se...
The paper introduces reViT, a recurrent vision transformer that uses a single Transformer block repeatedly to achieve the accuracy of a full-depth encoder while keeping inference FLOPs comparable. It achieves depth‑specific transformations by expressing each block’s feed‑forward network as a convex combination of a small shared expert bank, with a continuous depth coordinate guiding the mixture. Experiments on ImageNet‑1k and DINOv2 distillation show that reViT matches or surpasses state‑of‑the‑art models, reduces stored parameters by about 70%, and supports elastic‑depth deployment from a single checkpoint.
arXiv:2512.01675v3 Announce Type: replace Abstract: Text-to-image flow matching transformers degrade sharply in long-tail settings: tail-class outputs collapse in fidelity and diversity, limiting the...