REVIEW 21 cited by
Diffuse and Disperse: Image Generation with Representation Regularization
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
The development of diffusion-based generative models over the past decade has largely proceeded independently of progress in representation learning. These diffusion models typically rely on regression-based objectives and generally lack explicit regularization. In this work, we propose \textit{Dispersive Loss}, a simple plug-and-play regularizer that effectively improves diffusion-based generative models. Our loss function encourages internal representations to disperse in the hidden space, analogous to contrastive self-supervised learning, with the key distinction that it requires no positive sample pairs and therefore does not interfere with the sampling process used for regression. Compared to the recent method of representation alignment (REPA), our approach is self-contained and minimalist, requiring no pre-training, no additional parameters, and no external data. We evaluate Dispersive Loss on the ImageNet dataset across a range of models and report consistent improvements over widely used and strong baselines. We hope our work will help bridge the gap between generative modeling and representation learning.
Forward citations
Cited by 21 Pith papers
-
SPARE: Structural Parameter-Free Affinity Regularization for Flow Matching
A parameter-free regularizer that aligns intermediate token affinities to clean VAE latent affinities, including cross-image pairs, lowers FID on ImageNet with SiT backbones at matched training budgets.
-
Continuous Language Diffusion as a Decoder-Interface Problem
Continuous language diffusion works by entering high-margin decoder basins where frozen T5 embeddings recover 93-96% of native decisions and linear readouts reach 97.9% agreement, implying models should be evaluated a...
-
STREAM: Stochastic Riemannian Flow Matching with Anisotropic Decoder for Digital Histopathology Image Generation
STREAM applies stochastic Riemannian flow matching on VFM-derived unit hypersphere latents with a novel anisotropic decoder to achieve SOTA reconstruction and generation on breast and colorectal cancer histopathology ...
-
Genuine pair density wave order on the kagome lattice
A genuine primary pair-density-wave phase emerges as a competing ground state in a two-orbital kagome Hubbard model over a wide parameter range, driven by sublattice- and orbital-polarized Fermi pockets.
-
Mitigating Compounding Error via Video Representation Regularization
Compounding error in autoregressive video diffusion tracks effective-rank collapse of DiT hidden states, and representation regularization (SigReg/Unif) stabilizes long rollouts where data scaling does not.
-
Don't Settle at the Mode! Mitigating Diversity Collapse in Pretrained Flow Models via Feature Self-Guidance
Feature self-guidance disperses internal features of flow models during batch generation and applies manifold regularization to increase output diversity while preserving condition alignment.
-
DiffusionBench: On Holistic Evaluation of Diffusion Transformers
NanoGen unifies DiT training on ImageNet and T2I, reveals negative Pearson correlations (-0.377 to -0.580) in method rankings across metrics from 21 models, and motivates DiffusionBench for holistic evaluation.
-
HYDRA-X: Native Unified Multimodal Models with Holistic Visual Tokenizers
HYDRA-X presents the first unified multimodal model using a single ViT for holistic image-video tokenization, with ablations on attention and compression plus a latent-level editing improvement.
-
Generate in Reconstruction Space, Match in Semantic Space: Transport Geometry for One-Step Generation
Matching in semantic SSL feature space via Sinkhorn divergence enables effective one-step generation on ImageNet by inducing compact geometry for distribution matching, with training and evaluation features best kept ...
-
Improved Baselines with Representation Autoencoders
RAE v2 reaches gFID 1.06 on ImageNet-256 in 80 epochs by combining multi-layer encoder sums, complementary REPA targets, and free guidance via output reparameterization.
-
Stage-adaptive audio diffusion modeling
A semantic progress signal from SSL discrepancy slope enables three stage-aware mechanisms that improve training efficiency and performance in audio diffusion models over static baselines.
-
Continuous Adversarial Flow Models
Continuous adversarial flow models replace MSE in flow matching with adversarial training via a discriminator, improving guidance-free FID on ImageNet from 8.26 to 3.63 for SiT and similar gains for JiT and text-to-im...
-
MPDiT: Multi-Patch Global-to-Local Transformer Architecture For Efficient Flow Matching and Diffusion Model
MPDiT uses a hierarchical multi-patch design in transformers to lower computation in diffusion models by handling coarse global features first then fine local details, plus faster-converging embeddings.
-
Premier: Personalized Preference Modulation with Learnable User Embedding in Text-to-Image Generation
Premier learns user-specific embeddings to modulate text-to-image generation, outperforming prior methods on preference alignment, text consistency, and expert ratings even with limited history.
-
Self-transcendence: Is External Feature Guidance Indispensable for Accelerating Diffusion Transformer Training?
A self-supervised two-stage training method—VAE-latent feature alignment then feature-level classifier-free guidance—lets DiT models match or beat DINO-guided REPA training without any external feature extractor.
-
MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture
Training diffusion models on a mixture of higher-noise interpolations (MixFlow) improves generation FID across SiT, REPA, RAE and SD3.5, reaching ImageNet 256 gFID 1.43 after post-training.
-
Fitting Image Diffusion Models on Video Datasets
A shared-noise temporal consistency regularizer for image diffusion training accelerates convergence and lowers FID on the HandCo video dataset.
-
Amadeus: Autoregressive Model with Bidirectional Attribute Modelling for Symbolic Music
Amadeus generates symbolic music by autoregressively predicting note-level latents and decoding their attributes in parallel with a masked discrete diffusion model, yielding faster and more controllable generation tha...
-
Native3D: End-to-End 3D Scene Generation via Unified Mesh-Texture Modeling and Semantic Alignment
Native3D introduces a direct 3D scene generation method using unified mesh-texture representation and 3D REPA Loss for semantic alignment, claimed to outperform prior 2D-dependent approaches.
-
CoGenCast: A Coupled Autoregressive-Flow Generative Framework for Time Series Forecasting
CoGenCast couples a Qwen-based encoder-decoder with flow matching and reports strong MSE/MAE on ten time-series benchmarks.
-
Med-DisSeg: Dispersion-Driven Representation Learning for Fine-Grained Medical Image Segmentation
Med-DisSeg uses a dispersive loss on batch representations plus adaptive multi-scale decoding to achieve state-of-the-art fine-grained segmentation on five medical imaging datasets.
Discussion (0). Sign in to comment.