Data Unlearning via Inverse Distillation

arXiv:2609.36099 · cs.LG · Submitted 2026-09-28 · Read on arXiv

cs.LG

Submitted: 2026-09-28

Updated: 2026-09-28

Code: https://github.com/atong01/conditional-flow-matching

Terminology

• Distillation — 1× in this paper · explained in 2 episodes

Sources

• ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruning
• Unlearning for One-Step Generative Models via Unbalanced Optimal Transport
• EraseAnything++: Enabling Concept Erasure in Rectified Flow Transformers Leveraging Multi-Object Optimization
• Mean Flows for One-step Generative Modeling
• Denoising Diffusion Probabilistic Models
• Generator Matching: Generative modeling with arbitrary Markov processes
• Unlearning in Diffusion Models: A Unified Framework with KL Divergence and Likelihood Constraints
• Auto-Encoding Variational Bayes
• Erased or Dormant? Rethinking Concept Erasure Through Reversibility
• Unlearning or Concealment? A Critical Analysis and Evaluation Metrics for Unlearning in Diffusion Models
• ContinualFlow: Learning and Unlearning with Neural Flow Matching
• Improving and generalizing flow-based generative models with minibatch optimal transport

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