Research map / Image generation / Autoregressive and latent generation
Deep Automodulators
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add layer disentanglement loss L_j (ablation) | Automodulator · CelebA-HQ 256×256 | FID (mix): 52.83 -> 47.74 | Helped |
| Add layer disentanglement loss L_j (ablation) | Automodulator · CelebA-HQ 256×256 | PPL: 206.3 -> 210.0 | Hurt |
| Replace L1 reconstruction loss with robust loss d_rho, without L_j (ablation) | Automodulator · CelebA-HQ 256×256 | FID: 45.25 -> 36.20 | Helped |
| Replace L1 reconstruction loss with robust loss d_rho, without L_j (ablation) | Automodulator · CelebA-HQ 256×256 | FID (mix): 52.83 -> 43.53 | Helped |
| Replace L1 with d_rho without adding L_j (ablation) | Automodulator · CelebA-HQ 256×256 | FID (mix), compared with using both losses: 40.90 -> 43.53 | Hurt |
| Add L_j and replace L1 reconstruction loss with d_rho (ablation) | Automodulator · CelebA-HQ 256×256 | FID (mix): 52.83 -> 40.90 | Helped |
| Add learnable-scaled Gaussian noise before decoder activations | Automodulator · Image generation | FID | Helped |
Closest papers by meaning
- Adversarial Latent Autoencoders (CVPR 2020 · similarity 0.88)
- Swapping Autoencoder for Deep Image Manipulation (NeurIPS 2020 · similarity 0.88)
- A Primal-Dual link between GANs and Autoencoders (NeurIPS 2019 · similarity 0.88)
- Diffusion Autoencoders: Toward a Meaningful and Decodable Representation (CVPR 2022 · similarity 0.88)
- Autoencoding beyond pixels using a learned similarity metric (ICML 2016 · similarity 0.87)
- Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator (ICML 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.