Drift Variation Autoencoder: Unifying Generation and Representation Learning through Conditional Posterior Flow Matching
cs.LG, cs.AI
Submitted: 2026-08-25
Updated: 2026-08-25
License: http://creativecommons.org/licenses/by/4.0/
The gist: Stochastic masking, cropping, or modality removal makes deterministic reconstruction an incomplete target: one observation can admit many clean completions.
Terminology
Abstract
Stochastic masking, cropping, or modality removal makes deterministic reconstruction an incomplete target: one observation can admit many clean completions. This work takes the corresponding posterior P(X C) as the common statistical object for conditional generation and generatively sufficient representation learning. Drift Variation autoencoder trains a masked encoder Z=E(C) and a conditional flow decoder with one clean-prediction Flow Matching loss. The analysis first decomposes the ideal conditional KL into generator approximation and the representation deficiency I(X;C Z). It then derives orthogonal risk decompositions for conditional Flow Matching. For an affine Gaussian path, the clean-prediction representation gap is zero if and only if P(X Z)=P(X C). Thus the encoder-dependent excess clean-prediction risk induced by Flow Matching and the profiled ideal conditional KL have the same posterior-sufficient zero set, without being numerically equal objectives. An exact conditional field with a zero-noise endpoint then generates P(X Z) and hence P(X C) at a joint ideal optimum. The result extends to continuous multimodal product spaces when the complete modality tuple remains the Flow target for every observation mask. On CrossGeom-4, an 18-run controlled benchmark, observable factors have linear-probe R squared of 0.9990 - 0.9992, shuffling the joint model's encoder condition increases conditional error by 13.5 times - 15.7 times, and joint target attention reduces disagreement on an unobserved factor shared by two outputs by 90.1 - 92.8% relative to independent target decoders. Visible modalities are also generated and reconstructed, directly validating the full-tuple objective. Unconditional mode balance remains imperfect, delimiting the empirical claim to a controlled multimodal proof of concept.
Sources
- DINOv2: Learning Robust Visual Features without Supervision
- Score-Based Generative Modeling through Stochastic Differential Equations
- Flow Matching for Generative Modeling
- Flow Matching Guide and Code
- Zero-Flow Encoders
- Representation Learning with Diffusion Models
- Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think
- Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- High-Performance Self-Supervised Learning by Joint Training of Flow Matching
- Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis
- On Flow Matching KL Divergence
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