CBS partitions the diffusion timeline into segments of equal approximation burden via Dirichlet energy and trajectory acceleration monitors estimated by an auxiliary model, yielding higher synthesis quality at fixed per-step cost across SiT, JiT and UNet backbones.
Scaling laws for diffusion transformers
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
YoCausal benchmark shows video diffusion models detect the arrow of time but lack genuine causal understanding relative to humans.
Introduces a state-aligned latent actor-critic framework that lets diffusion models act as their own timestep-conditioned value functions for trajectory-level RL post-training and inference steering.
IAFS is a training-free iterative inference-time scaling framework that uses adaptive frequency-aware particle fusion to resolve the perception-fidelity conflict in diffusion super-resolution models, outperforming prior scaling strategies.
FID variance from training seeds is 3.2 times larger than from sampling seeds on hundreds of SiT models, with 1-2% coefficient of variation that barely shrinks with more compute, leading to a multi-seed evaluation protocol.
iTARFlow augments normalizing flows with diffusion-style iterative denoising during sampling while preserving end-to-end likelihood training, reaching competitive results on ImageNet 64/128/256.
Derives closed-form optimal loss for unified diffusion models, provides variance-controlled estimators, and shows improved diagnosis, training schedules, and power-law scaling after subtracting the optimal value.
Derives an asymptotic equivalent for the Representation Gap in equivariant diffusion models, showing it depends primarily on the intrinsic dimension of the task.
Continuous diffusion spoken language models follow scaling laws for loss and phoneme divergence and generate emotive multi-speaker speech at 16B scale, though long-form coherence stays difficult.
Turbo-GS accelerates 3D Gaussian Splatting training via dilated rendering of pixel subsets, convergence-aware Gaussian budget allocation, and combined positional-appearance error densification to enable faster 4K fitting with preserved or improved rendering quality.
Presents a single functional form for neural scaling that unifies multiple scaling dimensions and claims higher extrapolation accuracy than prior forms across diverse tasks and architectures.
Motif-Video 2B reaches 83.76% on VBench, outperforming a 14B-parameter model with 7x fewer parameters and far less training data through shared cross-attention and a three-part backbone.
citing papers explorer
-
Complexity-Balanced Diffusion Splitting
CBS partitions the diffusion timeline into segments of equal approximation burden via Dirichlet energy and trajectory acceleration monitors estimated by an auxiliary model, yielding higher synthesis quality at fixed per-step cost across SiT, JiT and UNet backbones.
-
YoCausal: How Far is Video Generation from World Model? A Causality Perspective
YoCausal benchmark shows video diffusion models detect the arrow of time but lack genuine causal understanding relative to humans.
-
Explicit Critic Guidance for Aligning Diffusion Models
Introduces a state-aligned latent actor-critic framework that lets diffusion models act as their own timestep-conditioned value functions for trajectory-level RL post-training and inference steering.
-
Iterative Inference-time Scaling with Adaptive Frequency Steering for Image Super-Resolution
IAFS is a training-free iterative inference-time scaling framework that uses adaptive frequency-aware particle fusion to resolve the perception-fidelity conflict in diffusion super-resolution models, outperforming prior scaling strategies.
-
The FID Lottery: Quantifying Hidden Randomness in Generative-Model Evaluation
FID variance from training seeds is 3.2 times larger than from sampling seeds on hundreds of SiT models, with 1-2% coefficient of variation that barely shrinks with more compute, leading to a multi-seed evaluation protocol.
-
Normalizing Flows with Iterative Denoising
iTARFlow augments normalizing flows with diffusion-style iterative denoising during sampling while preserving end-to-end likelihood training, reaching competitive results on ImageNet 64/128/256.
-
Diagnosing and Improving Diffusion Models by Estimating the Optimal Loss Value
Derives closed-form optimal loss for unified diffusion models, provides variance-controlled estimators, and shows improved diagnosis, training schedules, and power-law scaling after subtracting the optimal value.
-
Representation Gap: Explaining the Unreasonable Effectiveness of Neural Networks from a Geometric Perspective
Derives an asymptotic equivalent for the Representation Gap in equivariant diffusion models, showing it depends primarily on the intrinsic dimension of the task.
-
Scaling Properties of Continuous Diffusion Spoken Language Models
Continuous diffusion spoken language models follow scaling laws for loss and phoneme divergence and generate emotive multi-speaker speech at 16B scale, though long-form coherence stays difficult.
-
Turbo-GS: Accelerating 3D Gaussian Fitting for High-Quality Radiance Fields
Turbo-GS accelerates 3D Gaussian Splatting training via dilated rendering of pixel subsets, convergence-aware Gaussian budget allocation, and combined positional-appearance error densification to enable faster 4K fitting with preserved or improved rendering quality.
-
Unified Neural Scaling Laws
Presents a single functional form for neural scaling that unifies multiple scaling dimensions and claims higher extrapolation accuracy than prior forms across diverse tasks and architectures.
-
Motif-Video 2B: Technical Report
Motif-Video 2B reaches 83.76% on VBench, outperforming a 14B-parameter model with 7x fewer parameters and far less training data through shared cross-attention and a three-part backbone.