Data Attribution for Diffusion Models: Timestep-induced Bias in Influence Estimation (2401.09031v3)
Abstract: Data attribution methods trace model behavior back to its training dataset, offering an effective approach to better understand ''black-box'' neural networks. While prior research has established quantifiable links between model output and training data in diverse settings, interpreting diffusion model outputs in relation to training samples remains underexplored. In particular, diffusion models operate over a sequence of timesteps instead of instantaneous input-output relationships in previous contexts, posing a significant challenge to extend existing frameworks to diffusion models directly. Notably, we present Diffusion-TracIn that incorporates this temporal dynamics and observe that samples' loss gradient norms are highly dependent on timestep. This trend leads to a prominent bias in influence estimation, and is particularly noticeable for samples trained on large-norm-inducing timesteps, causing them to be generally influential. To mitigate this effect, we introduce Diffusion-ReTrac as a re-normalized adaptation that enables the retrieval of training samples more targeted to the test sample of interest, facilitating a localized measurement of influence and considerably more intuitive visualization. We demonstrate the efficacy of our approach through various evaluation metrics and auxiliary tasks, reducing the amount of generally influential samples to $\frac{1}{3}$ of its original quantity.
- Relatif: Identifying explanatory training samples via relative influence. In International Conference on Artificial Intelligence and Statistics, pages 1899–1909. PMLR, 2020.
- Influence functions in deep learning are fragile. arXiv preprint arXiv:2006.14651, 2020.
- Extracting training data from diffusion models. In 32nd USENIX Security Symposium (USENIX Security 23), pages 5253–5270, 2023.
- Training data attribution for diffusion models. arXiv preprint arXiv:2306.02174, 2023.
- Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34:8780–8794, 2021.
- Vitaly Feldman. Does learning require memorization? a short tale about a long tail. In Proceedings of the 52nd Annual ACM SIGACT Symposium on Theory of Computing, pages 954–959, 2020.
- The journey, not the destination: How data guides diffusion models. 2023.
- Generative adversarial networks. Communications of the ACM, 63(11):139–144, 2020.
- Identifying a training-set attack’s target using renormalized influence estimation. In Proceedings of the 2022 ACM SIGSAC Conference on Computer and Communications Security, pages 1367–1381, 2022.
- Prompt-to-prompt image editing with cross attention control. arXiv preprint arXiv:2208.01626, 2022.
- Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598, 2022.
- Denoising diffusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020.
- Imagen video: High definition video generation with diffusion models. arXiv preprint arXiv:2210.02303, 2022.
- Interpreting black box predictions using fisher kernels. In The 22nd International Conference on Artificial Intelligence and Statistics, pages 3382–3390. PMLR, 2019.
- Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013.
- Understanding black-box predictions via influence functions. In International conference on machine learning, pages 1885–1894. PMLR, 2017.
- Resolving training biases via influence-based data relabeling. In International Conference on Learning Representations, 2021.
- Understanding instance-based interpretability of variational auto-encoders. Advances in Neural Information Processing Systems, 34:2400–2412, 2021.
- Diffwave: A versatile diffusion model for audio synthesis. arXiv preprint arXiv:2009.09761, 2020.
- Learning multiple layers of features from tiny images. 2009.
- MNIST handwritten digit database. 2010. URL http://yann.lecun.com/exdb/mnist/.
- Diffusion-lm improves controllable text generation. Advances in Neural Information Processing Systems, 35:4328–4343, 2022.
- The artbench dataset: Benchmarking generative models with artworks. arXiv preprint arXiv:2206.11404, 2022.
- Measuring the effect of training data on deep learning predictions via randomized experiments. In International Conference on Machine Learning, pages 13468–13504. PMLR, 2022.
- Influence selection for active learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9274–9283, 2021.
- Trak: Attributing model behavior at scale. arXiv preprint arXiv:2303.14186, 2023.
- Estimating training data influence by tracing gradient descent. Advances in Neural Information Processing Systems, 33:19920–19930, 2020.
- High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022.
- Palette: Image-to-image diffusion models. In ACM SIGGRAPH 2022 Conference Proceedings, pages 1–10, 2022.
- Diffusion art or digital forgery? investigating data replication in diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6048–6058, 2023.
- Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502, 2020.
- Influence estimation for generative adversarial networks. arXiv preprint arXiv:2101.08367, 2021.
- Gerrit van den Burg and Chris Williams. On memorization in probabilistic deep generative models. Advances in Neural Information Processing Systems, 34:27916–27928, 2021.
- Representer point selection for explaining deep neural networks. Advances in neural information processing systems, 31, 2018.
- Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3836–3847, 2023.