From Pixels, Without Pre-training: Joint Generative and Self-Supervised Representation Learning in One Model
cs.LG, cs.AI, cs.CV, stat.ML
Submitted: 2026-10-05
Updated: 2026-10-05
Code: https://github.com/LTH14/rcg
Terminology
Sources
- PixelFlow: Pixel-Space Generative Models with Flow
- Big Self-Supervised Models are Strong Semi-Supervised Learners
- Flow Matching in Latent Space
- Unsupervised Representation Learning by Predicting Image Rotations
- Bootstrap your own latent: A new approach to self-supervised Learning
- Attention Residuals
- MAGE: MAsked Generative Encoder to Unify Representation Learning and Image Synthesis
- Latent Zoning Network: A Unified Principle for Generative Modeling, Representation Learning, and Classification
- Efficient Generative Model Training via Embedded Representation Warmup
- Conditional Generative Adversarial Nets
- Generating Images with Sparse Representations
- Robust Latent Matters: Boosting Image Generation with Sampling Error Synthesis
- GLU Variants Improve Transformer
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Representation Learning with Contrastive Predictive Coding
- Diffuse and Disperse: Image Generation with Representation Regularization
- PixNerd: Pixel Neural Field Diffusion
- FastCLIP: A Suite of Optimization Techniques to Accelerate CLIP Training with Limited Resources
- Repulsor: Accelerating Generative Modeling with a Contrastive Memory Bank
- Diffusion Transformers with Representation Autoencoders
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks