mlsanityData to 7 Oct 2026

Disentangling Regional Primitives for Image Generation

arXiv 2410.04421 ↗First posted 6 Oct 2024cs.CV cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

This paper explains a neural network for image generation from a new perspective, i.e., explaining representation structures for image generation. We propose a set of desirable properties to define the representation structure of a neural network for image generation, including feature completeness, spatial boundedness and consistency. These properties enable us to propose a method for disentangling primitive feature components from the intermediate-layer features, where each feature component generates a primitive regional pattern covering multiple image patches. In this way, the generation of the entire image can be explained as a superposition of these feature components. We prove that these feature components, which satisfy the feature completeness property and the linear additivity property (derived from the feature completeness, spatial boundedness, and consistency properties), can be computed as OR Harsanyi interaction. Experiments have verified the faithfulness of the disentangled primitive regional patterns.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 34 of its 44 references that are in the map. First posted on arXiv 6 Oct 2024.

Builds on

34 of 44 references.

References · 34

  1. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,690
  2. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  3. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  4. 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,580
  5. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  6. 2019StarGAN v2: Diverse Image Synthesis for Multiple Domainscited by 1,095
  7. 2021StyleCLIP: Text-Driven Manipulation of StyleGAN Imagerycited by 990
  8. 2019Interpreting the Latent Space of GANs for Semantic Face Editingcited by 691
  9. 2020NVAE: A Deep Hierarchical Variational Autoencodercited by 635
  10. 2020GANSpace: Discovering Interpretable GAN Controlscited by 612
  11. 2020Closed-Form Factorization of Latent Semantics in GANscited by 407
  12. 2020StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flowscited by 348
22 more
  1. 2022Diffusion Models already have a Semantic Latent Spacecited by 324
  2. 2020Unsupervised Discovery of Interpretable Directions in the GAN Latent Spacecited by 288
  3. 2024Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generationcited by 211
  4. 2023Self-Discovering Interpretable Diffusion Latent Directions for Responsible Text-to-Image Generationcited by 69
  5. 2021LatentCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directionscited by 52
  6. 2016Towards Better Analysis of Deep Convolutional Neural Networkscited by 51
  7. 2021Discovering and Explaining the Representation Bottleneck of DNNscited by 42
  8. 2021Defining and Quantifying the Emergence of Sparse Concepts in DNNscited by 36
  9. 2020Interpreting and Boosting Dropout from a Game-Theoretic Viewcited by 34
  10. 2023Does a Neural Network Really Encode Symbolic Concepts?cited by 27
  11. 2023Explaining Generalization Power of a DNN Using Interactive Conceptscited by 25
  12. 2024Defining and Extracting generalizable interaction primitives from DNNscited by 14
  13. 2024Towards the Dynamics of a DNN Learning Symbolic Interactionscited by 13
  14. 2022Proving Common Mechanisms Shared by Twelve Methods of Boosting Adversarial Transferabilitycited by 12
  15. 2023Can the Inference Logic of Large Language Models be Disentangled into Symbolic Concepts?cited by 11
  16. 2021Can We Faithfully Represent Masked States to Compute Shapley Values on a DNN?cited by 10
  17. 2022Exploring Gradient-based Multi-directional Controls in GANscited by 10
  18. 2023Bayesian Neural Networks Avoid Encoding Complex and Perturbation-Sensitive Conceptscited by 9
  19. 2023Explaining How a Neural Network Play the Go Game and Let People Learncited by 6
  20. 2024ASAP: Interpretable Analysis and Summarization of AI-generated Image Patterns at Scalecited by 5
  21. 2024Two-Phase Dynamics of Interactions Explains the Starting Point of a DNN Learning Over-Fitted Featurescited by 5
  22. 2024Layerwise Change of Knowledge in Neural Networkscited by 4

Built on by

Not cited yet.