ICML 2022 accepted papers on Image generation
20 of the 1,233 papers accepted at ICML 2022 (1.6%) are on Image generation, compared with 1.9% at ICML 2021. They fall into 5 topics.
See ICML 2022 in the interactive map
Submitting to ICML? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Autoregressive and latent generation · 7 papers
- Equivariant Diffusion for Molecule Generation in 3D (118 citations)
- Translatotron 2: High-quality direct speech-to-speech translation with voice preservation (20 citations)
- Guided-TTS: A Diffusion Model for Text-to-Speech via Classifier Guidance (19 citations)
- SkexGen: Autoregressive Generation of CAD Construction Sequences with Disentangled Codebooks (14 citations)
- Latent Diffusion Energy-Based Model for Interpretable Text Modelling (11 citations)
- A$^3$T: Alignment-Aware Acoustic and Text Pretraining for Speech Synthesis and Editing (1 citation)
- Bit Prioritization in Variational Autoencoders via Progressive Coding
GANs · 5 papers
- Region-Based Semantic Factorization in GANs (14 citations)
- Generative Cooperative Networks for Natural Language Generation (4 citations)
- A Neural Tangent Kernel Perspective of GANs (3 citations)
- Structure-preserving GANs (2 citations)
- Exploring and Exploiting Hubness Priors for High-Quality GAN Latent Sampling (1 citation)
Fast diffusion sampling · 4 papers
- Planning with Diffusion for Flexible Behavior Synthesis (60 citations)
- Estimating the Optimal Covariance with Imperfect Mean in Diffusion Probabilistic Models (18 citations)
- Maximum Likelihood Training for Score-based Diffusion ODEs by High Order Denoising Score Matching (11 citations)
- Soft Truncation: A Universal Training Technique of Score-based Diffusion Model for High Precision Score Estimation (11 citations)
