GUDA: Counterfactual Group-wise Training Data Attribution for Diffusion Models via Unlearning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914621182443520 |
|---|---|
| author | Murata, Naoki Takida, Yuhta Lai, Chieh-Hsin Uesaka, Toshimitsu Nguyen, Bac Ermon, Stefano Mitsufuji, Yuki |
| author_facet | Murata, Naoki Takida, Yuhta Lai, Chieh-Hsin Uesaka, Toshimitsu Nguyen, Bac Ermon, Stefano Mitsufuji, Yuki |
| contents | Training-data attribution for vision generative models aims to identify which training data influenced a given output. While most methods score individual examples, practitioners often need group-level answers (e.g., artistic styles or object classes). Group-wise attribution is counterfactual: how would a model's behavior on a generated sample change if a group were absent from training? A natural realization of this counterfactual is Leave-One-Group-Out (LOGO) retraining, which retrains the model with each group removed; however, it becomes computationally prohibitive as the number of groups grows. We propose GUDA (Group Unlearning-based Data Attribution) for diffusion models, which approximates each counterfactual model by applying machine unlearning to a shared full-data model instead of training from scratch. GUDA quantifies group influence using differences in a likelihood-based scoring rule (ELBO) between the full model and each unlearned counterfactual. Experiments on CIFAR-10 and artistic style attribution with Stable Diffusion show that GUDA identifies primary contributing groups more reliably than semantic similarity, gradient-based attribution, and instance-level unlearning approaches, while achieving ~100x speedup on CIFAR-10 over LOGO retraining. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22651 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | GUDA: Counterfactual Group-wise Training Data Attribution for Diffusion Models via Unlearning Murata, Naoki Takida, Yuhta Lai, Chieh-Hsin Uesaka, Toshimitsu Nguyen, Bac Ermon, Stefano Mitsufuji, Yuki Machine Learning Artificial Intelligence Training-data attribution for vision generative models aims to identify which training data influenced a given output. While most methods score individual examples, practitioners often need group-level answers (e.g., artistic styles or object classes). Group-wise attribution is counterfactual: how would a model's behavior on a generated sample change if a group were absent from training? A natural realization of this counterfactual is Leave-One-Group-Out (LOGO) retraining, which retrains the model with each group removed; however, it becomes computationally prohibitive as the number of groups grows. We propose GUDA (Group Unlearning-based Data Attribution) for diffusion models, which approximates each counterfactual model by applying machine unlearning to a shared full-data model instead of training from scratch. GUDA quantifies group influence using differences in a likelihood-based scoring rule (ELBO) between the full model and each unlearned counterfactual. Experiments on CIFAR-10 and artistic style attribution with Stable Diffusion show that GUDA identifies primary contributing groups more reliably than semantic similarity, gradient-based attribution, and instance-level unlearning approaches, while achieving ~100x speedup on CIFAR-10 over LOGO retraining. |
| title | GUDA: Counterfactual Group-wise Training Data Attribution for Diffusion Models via Unlearning |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2601.22651 |