Pith

open record

sign in

arxiv: 2112.10752 · v2 · pith:XFBNG3P6 · submitted 2021-12-20 · cs.CV

High-Resolution Image Synthesis with Latent Diffusion Models

Robin Rombach , Andreas Blattmann , Dominik Lorenz , Patrick Esser , Bj\"orn Ommer This is my paper

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 reserved 2026-05-11 21:56 UTCgrok-4.3pith:XFBNG3P6record.jsonopen to challenge →

classification cs.CV
keywords latent diffusion modelsimage synthesisdenoising diffusionautoencodersconditional generationimage inpaintingsuper-resolutioncross-attention
0
0 comments X

The pith

Diffusion models trained in the latent space of pretrained autoencoders generate high-resolution images with substantially lower computational cost than pixel-space versions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that diffusion models can be moved from raw pixel space into the compressed latent space of a fixed pretrained autoencoder. This shift preserves enough visual structure for high-fidelity synthesis while cutting the cost of training and sampling dramatically. Readers care because the same denoising process now supports conditioning via cross-attention layers, turning the model into a flexible generator for text, boxes, or masks without retraining. The result is practical high-resolution synthesis on ordinary hardware and new performance levels on inpainting.

Core claim

By applying the diffusion process to the latent representations of a pretrained autoencoder rather than to pixels, and by inserting cross-attention layers to accept arbitrary conditioning inputs, latent diffusion models reach a favorable trade-off between model capacity and perceptual fidelity while requiring far fewer resources than pixel-based diffusion models.

What carries the argument

The latent diffusion model (LDM), which runs the forward and reverse diffusion processes on the lower-dimensional latent codes produced by a fixed variational autoencoder and uses cross-attention to incorporate conditioning signals such as text or spatial layouts.

If this is right

  • Training and inference of powerful diffusion models become feasible on limited hardware while retaining visual quality.
  • High-resolution synthesis is performed directly in a convolutional manner without patch-wise processing.
  • Image inpainting reaches state-of-the-art results.
  • Unconditional generation, semantic scene synthesis, and super-resolution remain competitive with prior pixel-space methods.
  • Conditioning on text, bounding boxes, or other inputs is enabled without retraining the core model.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The separation of perceptual compression from the generative diffusion stage suggests similar latent-space training could be tested on other modalities once suitable autoencoders exist.
  • If the autoencoder is kept fixed, future improvements in autoencoder quality would immediately lift the upper bound on LDM fidelity without changing the diffusion architecture.
  • The approach implies that many existing pixel-based diffusion pipelines could be accelerated by first training a domain-specific autoencoder rather than scaling the diffusion model itself.

Load-bearing premise

The latent codes from the pretrained autoencoder already contain enough perceptual detail and spatial structure that the diffusion model can recover high-fidelity images without uncorrectable artifacts.

What would settle it

High-resolution outputs that consistently exhibit uncorrectable artifacts or visible loss of fine detail relative to pixel-based diffusion models of comparable training effort would show the assumption does not hold.

read the original abstract

By decomposing the image formation process into a sequential application of denoising autoencoders, diffusion models (DMs) achieve state-of-the-art synthesis results on image data and beyond. Additionally, their formulation allows for a guiding mechanism to control the image generation process without retraining. However, since these models typically operate directly in pixel space, optimization of powerful DMs often consumes hundreds of GPU days and inference is expensive due to sequential evaluations. To enable DM training on limited computational resources while retaining their quality and flexibility, we apply them in the latent space of powerful pretrained autoencoders. In contrast to previous work, training diffusion models on such a representation allows for the first time to reach a near-optimal point between complexity reduction and detail preservation, greatly boosting visual fidelity. By introducing cross-attention layers into the model architecture, we turn diffusion models into powerful and flexible generators for general conditioning inputs such as text or bounding boxes and high-resolution synthesis becomes possible in a convolutional manner. Our latent diffusion models (LDMs) achieve a new state of the art for image inpainting and highly competitive performance on various tasks, including unconditional image generation, semantic scene synthesis, and super-resolution, while significantly reducing computational requirements compared to pixel-based DMs. Code is available at https://github.com/CompVis/latent-diffusion .

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper claims that applying diffusion models in the latent space of pretrained autoencoders enables efficient high-resolution image synthesis. Latent diffusion models (LDMs) reduce spatial dimensions via a KL-regularized VAE (with downsampling factors f=4/8/16) while preserving detail, incorporate cross-attention for conditioning on text or bounding boxes, and achieve new state-of-the-art inpainting results along with competitive performance on unconditional generation, semantic synthesis, and super-resolution, all at substantially lower computational cost than pixel-space DMs. Public code is released.

Significance. If the results hold, this has high significance for making diffusion-based synthesis practical at high resolutions with limited resources. Strengths include the public code release, direct ablations on autoencoder factors, and quantitative FID/LPIPS tables on ImageNet, Places2, and ADE20K that support the efficiency and quality claims. The stress-test concern on latent representation fidelity does not land as a load-bearing issue, since the f=8 model empirically recovers high-frequency detail without uncorrectable artifacts and matches or exceeds pixel DM quality.

major comments (2)
  1. [Ablations on autoencoder downsampling factors] Ablations on autoencoder downsampling factors: the claim of reaching a 'near-optimal point' between complexity reduction and detail preservation for f=8 rests on FID comparisons, but the exact spatial cost reduction (stated as ~1/64) should be derived explicitly from the UNet channel dimensions and latent resolution to allow verification of the efficiency gain.
  2. [Cross-attention layers] Cross-attention for conditioning: while cross-attention enables flexible conditioning, the manuscript does not include an ablation against simpler conditioning mechanisms (e.g., concatenation or FiLM), which would isolate whether this architecture choice is necessary for the flexibility and high-resolution claims.
minor comments (2)
  1. [Abstract] The abstract's reference to 'hundreds of GPU days' for pixel-space DM optimization would be strengthened by citing the specific prior works being compared.
  2. [Methods] Notation for the latent variable z and the diffusion forward/reverse processes in latent space could be clarified with an explicit equation reference or diagram early in the methods section.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive assessment, the recommendation of minor revision, and the constructive comments on our efficiency claims and conditioning design. We address each major comment below and have incorporated revisions to improve clarity.

read point-by-point responses

  1. Referee: Ablations on autoencoder downsampling factors: the claim of reaching a 'near-optimal point' between complexity reduction and detail preservation for f=8 rests on FID comparisons, but the exact spatial cost reduction (stated as ~1/64) should be derived explicitly from the UNet channel dimensions and latent resolution to allow verification of the efficiency gain.

    Authors: We agree that an explicit derivation would strengthen the presentation. The ~1/64 factor follows directly from reducing the spatial resolution of the UNet input by f=8 in each dimension (latent size H/8 × W/8), which quadratically reduces the number of spatial operations. Accounting for the UNet channel schedule (starting at 320 channels with doubling in down-blocks), the overall computational cost of the diffusion process scales by this factor relative to pixel-space models. In the revised manuscript we will add a short derivation in Section 3.1 (or an appendix table) that computes the reduction from the exact latent resolution and channel dimensions, enabling straightforward verification. revision: yes

  2. Referee: Cross-attention for conditioning: while cross-attention enables flexible conditioning, the manuscript does not include an ablation against simpler conditioning mechanisms (e.g., concatenation or FiLM), which would isolate whether this architecture choice is necessary for the flexibility and high-resolution claims.

    Authors: We appreciate the suggestion. Cross-attention is chosen because it supports conditioning inputs of arbitrary length and structure (e.g., variable-length text token sequences or unordered sets of bounding-box embeddings) without requiring fixed-dimensional inputs, which concatenation or FiLM layers would necessitate. This flexibility is central to the high-resolution text-to-image and layout-to-image results. A full retraining ablation is outside the scope of a minor revision, but we will add a concise discussion paragraph in Section 3.2 explaining the architectural rationale and contrasting it with simpler alternatives, thereby addressing the concern without misrepresenting the design. revision: partial

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper proposes applying diffusion models in the latent space of a separately pretrained autoencoder, with the central claims of state-of-the-art inpainting and competitive performance on generation tasks supported by direct empirical ablations (e.g., downsampling factors f=4/8/16) and quantitative comparisons to pixel-space baselines on ImageNet, Places2, and ADE20K. No load-bearing step reduces a result or prediction to its own inputs by construction, fitted parameters renamed as outputs, or a self-citation chain; the autoencoder training and latent diffusion training are independent stages, and all performance assertions rest on measured metrics rather than theoretical closure.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that a pretrained autoencoder can compress images into a latent space that retains sufficient detail for diffusion-based generation; no free parameters are introduced in the abstract description, and no new entities are postulated.

axioms (1)
  • domain assumption Pretrained autoencoders produce latent representations that preserve perceptual details necessary for high-fidelity image synthesis.
    Invoked to justify operating diffusion in latent space rather than pixels.

pith-pipeline@v0.9.0 · 5539 in / 1297 out tokens · 47273 ms · 2026-05-11T21:56:11.949352+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 60 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Autoregressive Learning in Joint KL: Sharp Oracle Bounds and Lower Bounds

    cs.LG 2026-05 unverdicted novelty 8.0

    Joint KL yields horizon-free approximation but an information-theoretic lower bound of order Omega(H) for estimation error in autoregressive learning, with matching computationally efficient upper bounds.

  2. What Time Is It? How Data Geometry Makes Time Conditioning Optional for Flow Matching

    cs.LG 2026-05 unverdicted novelty 8.0

    Data geometry makes time identifiable from noisy interpolants at rate O(1/sqrt(d-k)), rendering the time-blindness gap asymptotically negligible relative to coupling variance.

  3. How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

    cs.LG 2026-04 unverdicted novelty 8.0

    FMRG reformulates guidance as deterministic optimal control, deriving a single-trajectory method using the flow map that matches or exceeds baselines on reward-guided generation and inverse problems with 3 NFEs at tex...

  4. Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution

    cs.CL 2023-09 unverdicted novelty 8.0

    Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.

  5. Tailor Made Embeddings for Quantum Machine Learning

    quant-ph 2026-06 unverdicted novelty 7.0

    A variational autoencoder learns quantum embeddings compressing ImageNet into 13 qubits and achieving 98.5% accuracy on MNIST 3-vs-5 classification with a quantum circuit, close to classical baselines and far above na...

  6. Ambient Diffusion Policy: Imitation Learning from Suboptimal Data in Robotics

    cs.RO 2026-06 unverdicted novelty 7.0

    Ambient Diffusion Policy enables better imitation learning from suboptimal robot data by leveraging spectral properties to restrict data usage to specific diffusion times.

  7. DRIFT: From Robustness Gaps to Invariance Manifolds for AI-Generated Image Detection

    cs.CV 2026-06 unverdicted novelty 7.0

    DRIFT learns a structured invariance manifold from real images via one-class supervision on decomposed robust and fragile subspaces of a frozen VFM to detect AI-generated images through margin violations.

  8. How Neural Losses Shape VAE Latents

    cs.LG 2026-05 unverdicted novelty 7.0

    Neural reconstruction losses in VAEs reduce latent information content and produce more isotropic latent geometries with even uncertainty distribution.

  9. DGLD: Domain-Gated Latent Diffusion for the Discovery of Novel Energetic Materials

    physics.chem-ph 2026-05 unverdicted novelty 7.0

    DGLD applies domain-gated latent diffusion with label-quality gating and multi-task guidance to discover 12 novel energetic material leads validated by DFT, outperforming SMILES-LSTM, SELFIES-GA, and REINVENT baseline...

  10. Constrained Code Generation with Discrete Diffusion

    cs.CL 2026-05 unverdicted novelty 7.0

    Constrained Diffusion for Code (CDC) integrates constraint satisfaction into the reverse denoising process of discrete diffusion models via constraint-aware operators that use optimization and program analysis to stee...

  11. Seeking the Unfamiliar but Memorable: Conceptual Creativity as Meta-Learning

    cs.LG 2026-05 unverdicted novelty 7.0

    Creativity is defined as meta-learning where a frozen diffusion creator optimizes candidates for rapid improvement by an adapting appraiser such as an autoencoder or CLIP adapter.

  12. CreFlow: Corrective Reflow for Sparse-Reward Embodied Video Diffusion RL

    cs.CV 2026-05 conditional novelty 7.0

    CreFlow combines LTL compositional rewards with credit-aware NFT and corrective reflow losses in online RL to improve embodied video diffusion models, raising downstream task success by 23.8 percentage points on eight...

  13. AuraMask: An Extensible Pipeline for Developing Aesthetic Anti-Facial Recognition Image Filters

    cs.CV 2026-05 conditional novelty 7.0

    AuraMask produces 40 aesthetic anti-facial recognition filters that match or exceed prior adversarial effectiveness and achieve significantly higher user acceptance in a 630-person study.

  14. Deep Dreams Are Made of This: Visualizing Monosemantic Features in Diffusion Models

    cs.LG 2026-05 unverdicted novelty 7.0

    LVO applies optimization-based feature visualization to latent diffusion models after disentangling their representations with sparse autoencoders, yielding recognizable concept images on a fine-tuned Stable Diffusion...

  15. How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

    cs.LG 2026-04 unverdicted novelty 7.0

    FMRG is a training-free, single-trajectory guidance method for flow models derived from optimal control that achieves strong reward alignment with only 3 NFEs.

  16. Oracle Noise: Faster Semantic Spherical Alignment for Interpretable Latent Optimization

    cs.CV 2026-04 unverdicted novelty 7.0

    Oracle Noise optimizes diffusion model noise on a Riemannian hypersphere guided by key prompt words to preserve the Gaussian prior, eliminate norm inflation, and achieve faster semantic alignment than Euclidean methods.

  17. $Z^2$-Sampling: Zero-Cost Zigzag Trajectories for Semantic Alignment in Diffusion Models

    cs.CV 2026-04 unverdicted novelty 7.0

    Z²-Sampling implicitly realizes zero-cost zigzag trajectories for curvature-aware semantic alignment in diffusion models by reducing multi-step paths via operator dualities and temporal caching while synthesizing a di...

  18. Flow of Truth: Proactive Temporal Forensics for Image-to-Video Generation

    cs.CV 2026-04 unverdicted novelty 7.0

    Flow of Truth introduces a learnable forensic template and template-guided flow module that follows pixel motion to enable temporal tracing in image-to-video generation.

  19. Flow of Truth: Proactive Temporal Forensics for Image-to-Video Generation

    cs.CV 2026-04 unverdicted novelty 7.0

    Flow of Truth is the first proactive temporal forensics framework for image-to-video generation that uses a learnable forensic template following pixel motion and a template-guided flow module to decouple motion from content.

  20. Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes

    cs.CV 2026-04 unverdicted novelty 7.0

    Text-to-3D models lose prompt sensitivity for out-of-distribution shapes due to sink traps but retain geometric diversity via unconditional priors, enabling a decoupled inversion method for robust editing.

  21. Advantage-Guided Diffusion for Model-Based Reinforcement Learning

    cs.AI 2026-04 unverdicted novelty 7.0

    Advantage-guided diffusion (SAG and EAG) steers sampling in diffusion world models to higher-advantage trajectories, enabling policy improvement and better sample efficiency on MuJoCo tasks.

  22. VASR: Variance-Aware Systematic Resampling for Reward-Guided Diffusion

    cs.AI 2026-04 unverdicted novelty 7.0

    FVD applies Fleming-Viot population dynamics to diffusion model sampling at inference time to reduce diversity collapse while improving reward alignment and FID scores.

  23. Drifting Fields are not Conservative

    cs.LG 2026-04 conditional novelty 7.0

    Drift fields in single-pass generative models are not conservative except for Gaussian kernels; a sharp kernel normalization makes them conservative for any radial kernel while noting that non-conservative fields offe...

  24. SEM-ROVER: Semantic Voxel-Guided Diffusion for Large-Scale Driving Scene Generation

    cs.CV 2026-04 unverdicted novelty 7.0

    SEM-ROVER generates large multiview-consistent 3D urban driving scenes via semantic-conditioned diffusion on Σ-Voxfield voxel grids with progressive outpainting and deferred rendering.

  25. Setting-Matched and Semantics-Scaled Benchmarking of One-Step Generative Models Against Multistep Diffusion and Flow Models

    cs.CV 2026-03 unverdicted novelty 7.0

    Matched benchmarking reveals FID misleads in few-step regimes under CFG, prompting CLIP-scaled and PickScore-scaled FID and IS variants for better semantic evaluation of one-step image generators.

  26. Latent Generative Solvers for Generalizable Long-Term Physics Simulation

    cs.AI 2026-02 unverdicted novelty 7.0

    LGS pretrained on 2.5M trajectories across 16 systems matches deterministic baselines at one step and halves 20-step error while using far less compute and adapting to held-out higher-resolution flows.

  27. Not All Denoising Steps Are Equal: Model Scheduling for Faster Masked Diffusion Language Models

    cs.LG 2026-02 unverdicted novelty 7.0

    Early and late denoising steps in masked diffusion LMs are robust to smaller-model replacement, enabling 17% FLOPs reduction with modest generative quality loss.

  28. Visual Diffusion Models are Geometric Solvers

    cs.CV 2025-10 unverdicted novelty 7.0

    Standard visual diffusion models operating in pixel space can approximate solutions to the inscribed square, Steiner tree, and simple polygon problems.

  29. VIPaint: Image Inpainting with Pre-Trained Diffusion Models via Variational Inference

    cs.CV 2024-11 unverdicted novelty 7.0

    VIPaint uses hierarchical variational inference to optimize a non-Gaussian Markov approximation of the diffusion posterior, enabling better inpainting and inverse problems with pre-trained and latent diffusion models.

  30. LAION-5B: An open large-scale dataset for training next generation image-text models

    cs.CV 2022-10 accept novelty 7.0

    LAION-5B is an openly released dataset of 5.85 billion CLIP-filtered image-text pairs that enables replication of foundational vision-language models.

  31. Hierarchical Text-Conditional Image Generation with CLIP Latents

    cs.CV 2022-04 accept novelty 7.0

    A hierarchical prior-decoder model using CLIP latents generates more diverse text-conditional images than direct methods while preserving photorealism and caption fidelity.

  32. Analysis-by-Proxy: Localization Signals in VLMs Operating as Condition Encoders

    cs.CV 2026-07 conditional novelty 6.0

    A lightweight Q-Former proxy trained on VLM hidden states reveals that localization signals peak in input-dependent intermediate layers, not the final layers used by standard editing pipelines.

  33. The Illusion of High Utility in Safety Alignment of Text-to-Image Diffusion Models

    cs.CV 2026-07 unverdicted novelty 6.0

    Safety-aligned T2I diffusion models exhibit semantic collapse in text embeddings causing TIFA drops; SAGE regularization restores structured utility while retaining safety.

  34. Chronos: A Physics-Informed Full-History Framework for Non-Markovian Long-Horizon Manipulation

    cs.RO 2026-06 unverdicted novelty 6.0

    Chronos elevates full observation history to the policy's latent state via selective SSM tokens and a Schrödinger-inspired acceleration bridge, achieving large gains on memory-dependent robot tasks with fewer parameters.

  35. SpatialFlow-GRPO: Where Spatial Credit Drives Image Editing

    cs.CV 2026-06 unverdicted novelty 6.0

    SpatialFlow-GRPO improves image editing quality by converting region-aware rewards into semantic-region-level optimization signals aligned with latent positions during policy updates.

  36. SpatialFlow-GRPO: Where Spatial Credit Drives Image Editing

    cs.CV 2026-06 unverdicted novelty 6.0

    SpatialFlow-GRPO adds region-level reward feedback and spatial alignment to Flow-GRPO-style RL for image editing, reporting gains on GEdit-Bench, ImgEdit-Bench, and a new MultiEditBench.

  37. Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models

    cs.CR 2026-06 unverdicted novelty 6.0

    A narrative survey that catalogs fifty papers on diffusion-based adversarial techniques across text, vision, and vision-language models, proposes a six-class taxonomy of diffusion roles plus a unified five-dimension e...

  38. One-Step Flow Matching for Generative Modeling of Path-Dependent Physical Fields

    cs.LG 2026-06 unverdicted novelty 6.0

    A one-step flow matching model using transformer in VAE latent space with non-Gaussian source and auxiliary networks generates accurate high-resolution path-dependent stress fields, achieving 6-7x CPU and ~100x GPU sp...

  39. Lighting-Consistent Object Transfer Across Radiance Fields

    cs.GR 2026-06 unverdicted novelty 6.0

    Diffusion-based per-view harmonization for lighting-consistent object transfer between 3DGS scenes, using heterogeneous training data and final 3D consolidation.

  40. Modality Forcing for Scalable Spatial Generation

    cs.CV 2026-06 unverdicted novelty 6.0

    Modality Forcing lets a single DiT produce image and depth outputs in any order after training on sparse real-world depth, with larger image-pretrained models yielding better depth accuracy and a 57% AbsRel reduction ...

  41. IDEAL: In-DEpth ALignment Makes A Discrete Representation AutoEncoder

    cs.CV 2026-06 unverdicted novelty 6.0

    IDEAL improves discrete representation autoencoders by jointly aligning quantized tokens with shallow and deep VFM features, reporting 0.61 rFID on ImageNet and 1.89 gFID for autoregressive image generation.

  42. Drifting Models for Surrogate Flow Modeling

    cs.LG 2026-06 unverdicted novelty 6.0

    A label-conditioned drifting model in VAE latent space matches diffusion accuracy for flow surrogates while running two orders of magnitude faster, with a spatial variant for unseen geometries.

  43. The Invisible Hand of Physics: When Video Diffusion Models Know More Than They Show

    cs.GR 2026-06 unverdicted novelty 6.0

    Physical plausibility is linearly decodable from diffusion transformer states in video models at 81.27% accuracy on IntPhys and InfLevel, absent from VAE latents and outperforming V-JEPA and VideoMAE.

  44. DiBS: Diffusion-Informed Branch Selection

    cs.AI 2026-06 unverdicted novelty 6.0

    DiBS ranks candidate values with a diffusion model inside a symbolic Sudoku solver to reduce nodes, backtracks, and long-tail search cost on the Royle 17-clue benchmark.

  45. SA-DTS: Semantic-Aware Digital Twin Synchronization over 6G Networks

    cs.ET 2026-06 unverdicted novelty 6.0

    SA-DTS achieves up to 94% bandwidth savings and 87% lower latency in digital twin synchronization by transmitting semantic features and reconstructing states with a partitioned knowledge graph.

  46. Generate in Reconstruction Space, Match in Semantic Space: Transport Geometry for One-Step Generation

    cs.LG 2026-05 unverdicted novelty 6.0

    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 ...

  47. Unlearning in Diffusion Models: A Unified Framework with KL Divergence and Likelihood Constraints

    cs.LG 2026-05 unverdicted novelty 6.0

    A constrained optimization framework for diffusion model unlearning via KL and likelihood constraints, with duality results and reported better retention-unlearning tradeoffs than weight-based baselines.

  48. Guidance Contrastive Token Credit Assignment for Discrete Policy Optimization

    cs.CV 2026-05 unverdicted novelty 6.0

    GCPO performs per-token credit assignment in discrete policy optimization by setting token advantages proportional to the difference in model predictions under positive versus negative prompts, outperforming GRPO and ...

  49. Continual Learning in Modern Hopfield Networks with an Application to Diffusion Models

    cs.LG 2026-05 unverdicted novelty 6.0

    Modern Hopfield energy identifies high-energy samples as more prone to intrinsic forgetting in continual learning, with effective energy-based replay validated in diffusion models.

  50. GenSBI: Generative Methods for Simulation-Based Inference in JAX

    cs.LG 2026-05 unverdicted novelty 6.0

    GenSBI delivers JAX-native implementations of generative SBI methods with transformer backbones and reports near-ideal calibration scores on standard benchmarks.

  51. Broken Memories: Detecting and Mitigating Memorization in Diffusion Models with Degraded Generations

    cs.CV 2026-05 unverdicted novelty 6.0

    Authors link memorization to internal instability in diffusion models via latent norms, propose step-wise detection and mitigation achieving AUC >0.999 and 0% memorization rate on Stable Diffusion 1.4.

  52. Broken Memories: Detecting and Mitigating Memorization in Diffusion Models with Degraded Generations

    cs.CV 2026-05 unverdicted novelty 6.0

    Memorization in diffusion models is detected via latent update norm instability and mitigated on-the-fly, yielding AUC over 0.999 and zero memorization rate on Stable Diffusion 1.4.

  53. UniVL: Unified Vision-Language Embedding for Spatially Grounded Contextual Image Generation

    cs.CV 2026-05 unverdicted novelty 6.0

    UniVL unifies vision and language into one mask-rendered input processed by an OCR backbone to condition diffusion models for spatially grounded image generation without a standalone text encoder.

  54. PaintCopilot: Modeling Painting as Autonomous Artistic Continuation

    cs.CV 2026-05 unverdicted novelty 6.0

    PaintCopilot models painting as an open-ended autoregressive process that predicts coherent brushstrokes from partial canvas observations using a ViT target predictor, flow-matching stroke generator, and VAE region sampler.

  55. Learning to Think in Physics: Breaking Shortcut Learning in Scientific Diffusion via Representation Alignment

    cs.LG 2026-05 unverdicted novelty 6.0

    REPA-P aligns intermediate representations in diffusion models with physical states using first-principles PDE residuals to accelerate convergence and boost out-of-distribution robustness on PDE tasks.

  56. Going PLACES: Participatory Localized Red Teaming for Text-to-Image Safety in the Global South

    cs.CY 2026-05 unverdicted novelty 6.0

    A participatory red-teaming project in the Global South created the PLACES dataset of 26k T2I failure examples that reveal unique cultural and linguistic harms missed by existing safety frameworks.

  57. A Distributional View for Visual Mechanistic Interpretability: KL-Minimal Soft-Constraint Principle

    cs.CV 2026-05 unverdicted novelty 6.0

    The work introduces a distributional view of visual mechanistic interpretability that casts the task as KL-minimal optimization and realizes it through a soft-constraint principle implemented with energy-guided diffus...

  58. MIRAGE: Robust multi-modal architectures translate fMRI-to-image models from vision to mental imagery

    q-bio.NC 2026-05 unverdicted novelty 6.0

    MIRAGE achieves state-of-the-art mental image reconstruction from fMRI on the NSD-Imagery benchmark by using a linear backbone with multi-modal text and image features fed to a diffusion model.

  59. Global Convergence of Sampling-Based Nonconvex Optimization through Diffusion-Style Smoothing

    cs.LG 2026-05 unverdicted novelty 6.0

    Recasts sampling-based nonconvex optimization as smoothed gradient descent to obtain non-asymptotic convergence guarantees and introduces the DIDA annealed algorithm that converges to the global optimum.

  60. Compositional Video Generation via Inference-Time Guidance

    cs.CV 2026-05 unverdicted novelty 6.0

    CVG improves compositional faithfulness in frozen text-to-video diffusion models by steering early denoising steps with gradients from a classifier trained on the model's own cross-attention features.

Reference graph

Works this paper leans on

109 extracted references · 109 canonical work pages · cited by 126 Pith papers · 16 internal anchors

  1. [1]

    NTIRE 2017 chal- lenge on single image super-resolution: Dataset and study

    Eirikur Agustsson and Radu Timofte. NTIRE 2017 chal- lenge on single image super-resolution: Dataset and study. In 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2017, Honolulu, HI, USA, July 21-26, 2017, pages 1122–1131. IEEE Com- puter Society, 2017. 1

    work page 2017

  2. [2]

    Wasserstein gan, 2017

    Martin Arjovsky, Soumith Chintala, and L ´eon Bottou. Wasserstein gan, 2017. 3

    work page 2017

  3. [3]

    Large scale GAN training for high fidelity natural image synthe- sis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale GAN training for high fidelity natural image synthe- sis. In Int. Conf. Learn. Represent. , 2019. 1, 2, 7, 8, 22, 28

    work page 2019

  4. [4]

    Holger Caesar, Jasper R. R. Uijlings, and Vittorio Ferrari. Coco-stuff: Thing and stuff classes in context. In 2018 IEEE Conference on Computer Vision and Pattern Recog- nition, CVPR 2018, Salt Lake City, UT, USA, June 18- 22, 2018, pages 1209–1218. Computer Vision Foundation / IEEE Computer Society, 2018. 7, 20, 22

    work page 2018

  5. [5]

    Extracting training data from large language models

    Nicholas Carlini, Florian Tramer, Eric Wallace, Matthew Jagielski, Ariel Herbert-V oss, Katherine Lee, Adam Roberts, Tom Brown, Dawn Song, Ulfar Erlingsson, et al. Extracting training data from large language models. In 30th USENIX Security Symposium (USENIX Security 21) , pages 2633–2650, 2021. 9

    work page 2021

  6. [6]

    Generative pre- training from pixels

    Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Hee- woo Jun, David Luan, and Ilya Sutskever. Generative pre- training from pixels. In ICML, volume 119 of Proceedings of Machine Learning Research, pages 1691–1703. PMLR,

    work page

  7. [7]

    Weiss, Mo- hammad Norouzi, and William Chan

    Nanxin Chen, Yu Zhang, Heiga Zen, Ron J. Weiss, Mo- hammad Norouzi, and William Chan. Wavegrad: Estimat- ing gradients for waveform generation. In ICLR. OpenRe- view.net, 2021. 1

    work page 2021

  8. [8]

    Fast fourier convolu- tion

    Lu Chi, Borui Jiang, and Yadong Mu. Fast fourier convolu- tion. In NeurIPS, 2020. 8

    work page 2020

  9. [9]

    CHILD, Very deep vaes, arXiv preprint arXiv:2011.10650, (2020), https://doi.org/10.48550/ arXiv.2011.10650

    Rewon Child. Very deep vaes generalize autoregressive models and can outperform them on images. CoRR, abs/2011.10650, 2020. 3

    work page arXiv 2011

  10. [10]

    Generating Long Sequences with Sparse Transformers

    Rewon Child, Scott Gray, Alec Radford, and Ilya Sutskever. Generating long sequences with sparse transformers. CoRR, abs/1904.10509, 2019. 3

    work page internal anchor Pith review Pith/arXiv arXiv 1904

  11. [11]

    Bin Dai and David P. Wipf. Diagnosing and enhancing V AE models. In ICLR (Poster). OpenReview.net, 2019. 2, 3

    work page 2019

  12. [12]

    Imagenet: A large-scale hierarchical im- age database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Fei-Fei Li. Imagenet: A large-scale hierarchical im- age database. In CVPR, pages 248–255. IEEE Computer Society, 2009. 1, 5, 7, 22

    work page 2009

  13. [13]

    Ethical considerations of generative ai

    Emily Denton. Ethical considerations of generative ai. AI for Content Creation Workshop, CVPR, 2021. 9

    work page 2021

  14. [14]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: pre-training of deep bidirec- tional transformers for language understanding. CoRR, abs/1810.04805, 2018. 7

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  15. [15]

    Diffusion Models Beat GANs on Image Synthesis

    Prafulla Dhariwal and Alex Nichol. Diffusion models beat gans on image synthesis. CoRR, abs/2105.05233, 2021. 1, 2, 3, 4, 6, 7, 8, 18, 22, 25, 26, 28

    work page internal anchor Pith review arXiv 2021

  16. [16]

    Musings on typicality, 2020

    Sander Dieleman. Musings on typicality, 2020. 1, 3

    work page 2020

  17. [17]

    Advances in Neural Information Processing Systems (NeurIPS) , year=

    Ming Ding, Zhuoyi Yang, Wenyi Hong, Wendi Zheng, Chang Zhou, Da Yin, Junyang Lin, Xu Zou, Zhou Shao, Hongxia Yang, and Jie Tang. Cogview: Mastering text-to- image generation via transformers. CoRR, abs/2105.13290,

    work page arXiv

  18. [18]

    Nice: Non-linear independent components estimation, 2015

    Laurent Dinh, David Krueger, and Yoshua Bengio. Nice: Non-linear independent components estimation, 2015. 3

    work page 2015

  19. [19]

    Density estimation using real NVP

    Laurent Dinh, Jascha Sohl-Dickstein, and Samy Ben- gio. Density estimation using real NVP. In 5th Inter- national Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net, 2017. 1, 3

    work page 2017

  20. [20]

    Generating images with perceptual similarity metrics based on deep networks

    Alexey Dosovitskiy and Thomas Brox. Generating images with perceptual similarity metrics based on deep networks. In Daniel D. Lee, Masashi Sugiyama, Ulrike von Luxburg, Isabelle Guyon, and Roman Garnett, editors, Adv. Neural Inform. Process. Syst., pages 658–666, 2016. 3

    work page 2016

  21. [21]

    Imagebart: Bidirectional context with multi- nomial diffusion for autoregressive image synthesis.CoRR, abs/2108.08827, 2021

    Patrick Esser, Robin Rombach, Andreas Blattmann, and Bj¨orn Ommer. Imagebart: Bidirectional context with multi- nomial diffusion for autoregressive image synthesis.CoRR, abs/2108.08827, 2021. 6, 7, 22

    work page arXiv 2021

  22. [22]

    A note on data biases in generative models

    Patrick Esser, Robin Rombach, and Bj ¨orn Ommer. A note on data biases in generative models. arXiv preprint arXiv:2012.02516, 2020. 9

    work page arXiv 2012

  23. [23]

    Taming Transformers for High-Resolution Image Synthesis

    Patrick Esser, Robin Rombach, and Bj ¨orn Ommer. Taming transformers for high-resolution image synthesis. CoRR, abs/2012.09841, 2020. 2, 3, 4, 6, 7, 21, 22, 29, 34, 36

    work page arXiv 2012

  24. [24]

    Sex, lies, and videotape: Deep fakes and free speech delusions

    Mary Anne Franks and Ari Ezra Waldman. Sex, lies, and videotape: Deep fakes and free speech delusions. Md. L. Rev., 78:892, 2018. 9

    work page 2018

  25. [25]

    Soros, and Olaf Witkowski

    Kevin Frans, Lisa B. Soros, and Olaf Witkowski. Clipdraw: Exploring text-to-drawing synthesis through language- image encoders. ArXiv, abs/2106.14843, 2021. 3

    work page arXiv 2021

  26. [26]

    Make-a-scene: Scene-based text-to-image generation with human priors.ArXiv, abs/2203.13131, 2022

    Oran Gafni, Adam Polyak, Oron Ashual, Shelly Sheynin, Devi Parikh, and Yaniv Taigman. Make-a-scene: Scene- based text-to-image generation with human priors. CoRR, abs/2203.13131, 2022. 6, 7, 16

    work page arXiv 2022

  27. [27]

    Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron C

    Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron C. Courville, and Yoshua Bengio. Generative adversarial networks. CoRR, 2014. 1, 2

    work page 2014

  28. [28]

    Improved training of wasserstein gans, 2017

    Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, and Aaron Courville. Improved training of wasserstein gans, 2017. 3

    work page 2017

  29. [29]

    Gans trained by a two time-scale update rule converge to a local nash equi- librium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equi- librium. In Adv. Neural Inform. Process. Syst., pages 6626– 6637, 2017. 1, 5, 26

    work page 2017

  30. [30]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. In NeurIPS, 2020. 1, 2, 3, 4, 6, 17

    work page 2020

  31. [31]

    Cascaded diffusion models for high fidelity image generation,

    Jonathan Ho, Chitwan Saharia, William Chan, David J. Fleet, Mohammad Norouzi, and Tim Salimans. Cascaded diffusion models for high fidelity image generation.CoRR, abs/2106.15282, 2021. 1, 3, 22 10

    work page arXiv 2021

  32. [32]

    Classifier-free diffusion guidance

    Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. In NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications, 2021. 6, 7, 16, 22, 28, 37, 38

    work page 2021

  33. [33]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-image translation with conditional adver- sarial networks. In CVPR, pages 5967–5976. IEEE Com- puter Society, 2017. 3, 4

    work page 2017

  34. [34]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-image translation with conditional adver- sarial networks. 2017 IEEE Conference on Computer Vi- sion and Pattern Recognition (CVPR) , pages 5967–5976,

    work page 2017

  35. [35]

    Perceiver IO: A General Architecture for Structured Inputs & Outputs

    Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier J. H ´enaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, and Jo ˜ao Carreira. Perceiver IO: A general architecture for structured inputs &outputs. CoRR, abs/2107.14795, 2021. 4

    work page internal anchor Pith review arXiv 2021

  36. [36]

    Perceiver: General perception with iterative attention

    Andrew Jaegle, Felix Gimeno, Andy Brock, Oriol Vinyals, Andrew Zisserman, and Jo ˜ao Carreira. Perceiver: General perception with iterative attention. In Marina Meila and Tong Zhang, editors, Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18-24 July 2021, Virtual Event, volume 139 ofProceedings of Machine Learning Researc...

    work page 2021

  37. [37]

    High- resolution complex scene synthesis with transformers

    Manuel Jahn, Robin Rombach, and Bj ¨orn Ommer. High- resolution complex scene synthesis with transformers. CoRR, abs/2105.06458, 2021. 20, 22, 27

    work page arXiv 2021

  38. [38]

    Imperfect Ima- GANation: Implications of GANs exacerbating biases on facial data augmentation and Snapchat selfie lenses.arXiv preprint arXiv:2001.09528, 2020

    Niharika Jain, Alberto Olmo, Sailik Sengupta, Lydia Manikonda, and Subbarao Kambhampati. Imperfect ima- ganation: Implications of gans exacerbating biases on fa- cial data augmentation and snapchat selfie lenses. arXiv preprint arXiv:2001.09528, 2020. 9

    work page arXiv 2001

  39. [39]

    Progressive Growing of GANs for Improved Quality, Stability, and Variation

    Tero Karras, Timo Aila, Samuli Laine, and Jaakko Lehti- nen. Progressive growing of gans for improved quality, sta- bility, and variation. CoRR, abs/1710.10196, 2017. 5, 6

    work page internal anchor Pith review arXiv 2017

  40. [40]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In IEEE Conf. Comput. Vis. Pattern Recog. , pages 4401– 4410, 2019. 1

    work page 2019

  41. [41]

    Karras, S

    T. Karras, S. Laine, and T. Aila. A style-based gener- ator architecture for generative adversarial networks. In 2019 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), 2019. 5, 6

    work page 2019

  42. [42]

    Karras, S

    Tero Karras, Samuli Laine, Miika Aittala, Janne Hellsten, Jaakko Lehtinen, and Timo Aila. Analyzing and improv- ing the image quality of stylegan. CoRR, abs/1912.04958,

    work page arXiv 1912

  43. [43]

    Soft truncation: A universal training technique of score-based diffusion model for high precision score estimation

    Dongjun Kim, Seungjae Shin, Kyungwoo Song, Wanmo Kang, and Il-Chul Moon. Score matching model for un- bounded data score. CoRR, abs/2106.05527, 2021. 6

    work page arXiv 2021

  44. [44]

    Glow: Generative flow with invertible 1x1 convolutions

    Durk P Kingma and Prafulla Dhariwal. Glow: Generative flow with invertible 1x1 convolutions. In S. Bengio, H. Wal- lach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors, Advances in Neural Information Process- ing Systems, 2018. 3

    work page 2018

  45. [45]

    Variational diffusion models.arXiv preprint arXiv:2107.00630, 2,

    Diederik P. Kingma, Tim Salimans, Ben Poole, and Jonathan Ho. Variational diffusion models. CoRR, abs/2107.00630, 2021. 1, 3, 16

    work page arXiv 2021

  46. [46]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. Auto-Encoding Vari- ational Bayes. In 2nd International Conference on Learn- ing Representations, ICLR, 2014. 1, 3, 4, 29

    work page 2014

  47. [47]

    On fast sampling of diffusion probabilistic models,

    Zhifeng Kong and Wei Ping. On fast sampling of diffusion probabilistic models. CoRR, abs/2106.00132, 2021. 3

    work page arXiv 2021

  48. [48]

    Diffwave: A versatile diffusion model for audio synthesis

    Zhifeng Kong, Wei Ping, Jiaji Huang, Kexin Zhao, and Bryan Catanzaro. Diffwave: A versatile diffusion model for audio synthesis. In ICLR. OpenReview.net, 2021. 1

    work page 2021

  49. [49]

    Alina Kuznetsova, Hassan Rom, Neil Alldrin, Jasper R. R. Uijlings, Ivan Krasin, Jordi Pont-Tuset, Shahab Kamali, Stefan Popov, Matteo Malloci, Tom Duerig, and Vittorio Ferrari. The open images dataset V4: unified image classi- fication, object detection, and visual relationship detection at scale. CoRR, abs/1811.00982, 2018. 7, 20, 22

    work page arXiv 2018

  50. [50]

    Kynkäänniemi, T

    Tuomas Kynk ¨a¨anniemi, Tero Karras, Samuli Laine, Jaakko Lehtinen, and Timo Aila. Improved precision and re- call metric for assessing generative models. CoRR, abs/1904.06991, 2019. 5, 26

    work page arXiv 1904

  51. [51]

    Microsoft COCO: Common Objects in Context

    Tsung-Yi Lin, Michael Maire, Serge J. Belongie, Lubomir D. Bourdev, Ross B. Girshick, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C. Lawrence Zit- nick. Microsoft COCO: common objects in context. CoRR, abs/1405.0312, 2014. 6, 7, 27

    work page internal anchor Pith review arXiv 2014

  52. [52]

    Region-wise generative adversarial imageinpainting for large missing ar- eas

    Yuqing Ma, Xianglong Liu, Shihao Bai, Le-Yi Wang, Ais- han Liu, Dacheng Tao, and Edwin Hancock. Region-wise generative adversarial imageinpainting for large missing ar- eas. ArXiv, abs/1909.12507, 2019. 9

    work page arXiv 1909

  53. [53]

    SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations

    Chenlin Meng, Yang Song, Jiaming Song, Jiajun Wu, Jun- Yan Zhu, and Stefano Ermon. Sdedit: Image synthesis and editing with stochastic differential equations. CoRR, abs/2108.01073, 2021. 1

    work page internal anchor Pith review arXiv 2021

  54. [54]

    Which Training Methods for GANs do actually Converge?

    Lars M. Mescheder. On the convergence properties of GAN training. CoRR, abs/1801.04406, 2018. 3

    work page Pith review arXiv 2018

  55. [55]

    Unrolled generative adversarial networks

    Luke Metz, Ben Poole, David Pfau, and Jascha Sohl- Dickstein. Unrolled generative adversarial networks. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net, 2017. 3

    work page 2017

  56. [56]

    Conditional Generative Adversarial Nets

    Mehdi Mirza and Simon Osindero. Conditional generative adversarial nets. CoRR, abs/1411.1784, 2014. 4

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  57. [57]

    Engel, Curtis Hawthorne, and Ian Simon

    Gautam Mittal, Jesse H. Engel, Curtis Hawthorne, and Ian Simon. Symbolic music generation with diffusion models. CoRR, abs/2103.16091, 2021. 1

    work page arXiv 2021

  58. [58]

    EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning

    Kamyar Nazeri, Eric Ng, Tony Joseph, Faisal Z. Qureshi, and Mehran Ebrahimi. Edgeconnect: Generative im- age inpainting with adversarial edge learning. ArXiv, abs/1901.00212, 2019. 9

    work page Pith review arXiv 1901

  59. [59]

    GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. GLIDE: towards photorealistic image genera- tion and editing with text-guided diffusion models. CoRR, abs/2112.10741, 2021. 6, 7, 16

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  60. [60]

    Saßnick and C

    Anton Obukhov, Maximilian Seitzer, Po-Wei Wu, Se- men Zhydenko, Jonathan Kyl, and Elvis Yu-Jing Lin. 11 High-fidelity performance metrics for generative models in pytorch, 2020. Version: 0.3.0, DOI: 10.5281/zen- odo.4957738. 26, 27

    work page doi:10.5281/zen- 2020

  61. [61]

    Semantic image synthesis with spatially-adaptive normalization

    Taesung Park, Ming-Yu Liu, Ting-Chun Wang, and Jun- Yan Zhu. Semantic image synthesis with spatially-adaptive normalization. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019. 4, 7

    work page 2019

  62. [62]

    Semantic image synthesis with spatially-adaptive normalization

    Taesung Park, Ming-Yu Liu, Ting-Chun Wang, and Jun- Yan Zhu. Semantic image synthesis with spatially-adaptive normalization. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR), June 2019. 22

    work page 2019

  63. [63]

    Dual contradistinctive generative autoencoder

    Gaurav Parmar, Dacheng Li, Kwonjoon Lee, and Zhuowen Tu. Dual contradistinctive generative autoencoder. InIEEE Conference on Computer Vision and Pattern Recognition, CVPR 2021, virtual, June 19-25, 2021 , pages 823–832. Computer Vision Foundation / IEEE, 2021. 6

    work page 2021

  64. [64]

    arXiv preprint arXiv:2104.11222 , year=

    Gaurav Parmar, Richard Zhang, and Jun-Yan Zhu. On buggy resizing libraries and surprising subtleties in fid cal- culation. arXiv preprint arXiv:2104.11222, 2021. 26

    work page arXiv 2021

  65. [65]

    Carbon Emissions and Large Neural Network Training

    David A. Patterson, Joseph Gonzalez, Quoc V . Le, Chen Liang, Lluis-Miquel Munguia, Daniel Rothchild, David R. So, Maud Texier, and Jeff Dean. Carbon emissions and large neural network training. CoRR, abs/2104.10350,

    work page internal anchor Pith review arXiv

  66. [66]

    Zero-Shot Text-to-Image Generation

    Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea V oss, Alec Radford, Mark Chen, and Ilya Sutskever. Zero-shot text-to-image generation. CoRR, abs/2102.12092, 2021. 1, 2, 3, 4, 7, 21, 27

    work page internal anchor Pith review arXiv 2021

  67. [67]

    Gen- erating diverse high-fidelity images with VQ-V AE-2

    Ali Razavi, A ¨aron van den Oord, and Oriol Vinyals. Gen- erating diverse high-fidelity images with VQ-V AE-2. In NeurIPS, pages 14837–14847, 2019. 1, 2, 3, 22

    work page 2019

  68. [68]

    Reed, Zeynep Akata, Xinchen Yan, Lajanugen Lo- geswaran, Bernt Schiele, and Honglak Lee

    Scott E. Reed, Zeynep Akata, Xinchen Yan, Lajanugen Lo- geswaran, Bernt Schiele, and Honglak Lee. Generative ad- versarial text to image synthesis. In ICML, 2016. 4

    work page 2016

  69. [69]

    Stochastic backpropagation and approximate in- ference in deep generative models

    Danilo Jimenez Rezende, Shakir Mohamed, and Daan Wierstra. Stochastic backpropagation and approximate in- ference in deep generative models. In Proceedings of the 31st International Conference on International Conference on Machine Learning, ICML, 2014. 1, 4, 29

    work page 2014

  70. [70]

    Network-to-network translation with conditional invertible neural networks

    Robin Rombach, Patrick Esser, and Bj ¨orn Ommer. Network-to-network translation with conditional invertible neural networks. In NeurIPS, 2020. 3

    work page 2020

  71. [71]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In MICCAI (3), volume 9351 of Lecture Notes in Computer Science, pages 234–241. Springer, 2015. 2, 3, 4

    work page 2015

  72. [72]

    Saharia, J

    Chitwan Saharia, Jonathan Ho, William Chan, Tim Sal- imans, David J. Fleet, and Mohammad Norouzi. Im- age super-resolution via iterative refinement. CoRR, abs/2104.07636, 2021. 1, 4, 8, 16, 22, 23, 27

    work page arXiv 2021

  73. [73]

    Tim Salimans, Andrej Karpathy, Xi Chen, and Diederik P. Kingma. Pixelcnn++: Improving the pixelcnn with dis- cretized logistic mixture likelihood and other modifications. CoRR, abs/1701.05517, 2017. 1, 3

    work page Pith review arXiv 2017

  74. [74]

    NVIDIA Developer Blog

    Dave Salvator. NVIDIA Developer Blog. https : / / developer . nvidia . com / blog / getting - immediate- speedups- with- a100- tf32, 2020. 28

    work page 2020

  75. [75]

    Noise estim ation for generative diffusion models

    Robin San-Roman, Eliya Nachmani, and Lior Wolf. Noise estimation for generative diffusion models. CoRR, abs/2104.02600, 2021. 3

    work page arXiv 2021

  76. [76]

    Projected gans converge faster

    Axel Sauer, Kashyap Chitta, Jens M ¨uller, and An- dreas Geiger. Projected gans converge faster. CoRR, abs/2111.01007, 2021. 6

    work page arXiv 2021

  77. [77]

    A u- net based discriminator for generative adversarial networks

    Edgar Sch ¨onfeld, Bernt Schiele, and Anna Khoreva. A u- net based discriminator for generative adversarial networks. In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2020, Seattle, WA, USA, June 13-19, 2020, pages 8204–8213. Computer Vision Founda- tion / IEEE, 2020. 6

    work page 2020

  78. [78]

    Laion- 400m: Open dataset of clip-filtered 400 million image-text pairs, 2021

    Christoph Schuhmann, Richard Vencu, Romain Beaumont, Robert Kaczmarczyk, Clayton Mullis, Aarush Katta, Theo Coombes, Jenia Jitsev, and Aran Komatsuzaki. Laion- 400m: Open dataset of clip-filtered 400 million image-text pairs, 2021. 6, 7

    work page 2021

  79. [79]

    Very deep con- volutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep con- volutional networks for large-scale image recognition. In Yoshua Bengio and Yann LeCun, editors, Int. Conf. Learn. Represent., 2015. 29, 43, 44, 45

    work page 2015

  80. [80]

    D2C: diffusion-denoising models for few-shot con- ditional generation

    Abhishek Sinha, Jiaming Song, Chenlin Meng, and Stefano Ermon. D2C: diffusion-denoising models for few-shot con- ditional generation. CoRR, abs/2106.06819, 2021. 3

    work page arXiv 2021

Showing first 80 references.