Data Attribution for Text-to-Image Models by Unlearning Synthesized Images

The goal of data attribution for text-to-image models is to identify the training images that most influence the generation of a new image. Influence is defined such that, for a given output, if a model is retrained from scratch without the most influential images, the model would fail to reproduce the same output. Unfortunately, directly searching for these influential images is computationally infeasible, since it would require repeatedly retraining models from scratch. In our work, we propose an efficient data attribution method by simulating unlearning the synthesized image. We achieve this by increasing the training loss on the output image, without catastrophic forgetting of other, unrelated concepts. We then identify training images with significant loss deviations after the unlearning process and label these as influential. We evaluate our method with a computationally intensive but"gold-standard"retraining from scratch and demonstrate our method's advantages over previous methods.

Paper

Similar papers

Peer review

Reviewer YhLa6/10 · confidence 3/52024-07-13

Summary

The authors propose a new approach for data attribution for text-to-image models that utilizes machine unlearning. They then perform multiple experiments to demonstrate that the proposed method is competitive to other methods.

Strengths

- The paper is generally well-written, clearly structured, and easy to follow. - The problem of data attribution in text-to-image models is interesting. - Multiple experiments are conducted to verify the viewpoints.

Weaknesses

- This is not really a weakness, but I wonder if the leave-K-out model can still generate a target image just under a different prompt. Perhaps a metric that see how hard it is to perform Textual Inversion [1] can be used? References: [1] Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patashnik, Amit H. Bermano, Gal Chechik, Daniel Cohen-Or, “An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion”

Questions

- See weaknesses.

Rating

6

Confidence

3

Soundness

3

Presentation

3

Contribution

3

Limitations

- The authors have adequately addressed the limitations.

Reviewer UBLR6/10 · confidence 3/52024-07-13

Summary

The paper discusses how we can identify influential examples by unlearning the synthesized images in the text-to-image generation models. Unlearning synthesized images leads to increase in the loss for the most influential training examples for the generation of synthesized images. The paper relies on a well-known regularization for Fisher information in order to avoid catastrophic forgetting. The paper evaluates image generation quality two ways: 1) Retraining the model from scratch after removing the influential examples from the training set. A Poor synthetic image generation quality is seen as a proxy for identifying most influential training examples. 2) Using ground truth attribution provided in a Customized Model Benchmark.

Strengths

+ The authors did a great job telling the story, motivating the problem and discussing related work. + The evaluation metrics and benchmarks are well-described. The authors show the effectiveness of their approach both qualitatively and quantitatively.

Weaknesses

+ It is unclear how robust the synthetic image generation for a given caption is. It would be good to estimate the sensitivity to the subtle changes in the caption to the generation quality. + Overall theoretical justification is not very clear. It is not clear how unlearning affects the overall utility of the model. + Figure 2 mentions that their approach is qualitatively better than DINO and JourneyTRAK, but how can we tell that it is quantitatively better ? It might be that the proposed method is better for some examples and worse for others. In the implementation details section, line 212, the paper suddenly starts mentioning DDPM loss without explaining what it is and how it is applied in their method. In the methods section there is no mention of it but instead the paper discusses a different loss, EWC loss. This is a bit confusing. + On Figures 3 and 2 DINO is listed as a baseline against which the authors evaluated the proposed method. It is, however, not an influence function's approximation or unlearning method. It seems that the authors used it for Image similarity. At the same time they compare it also against Journey TRAK which is an influence approximation approach. This is a bit confusing. + Figure 4: where exactly are the error bars in the plot ?

Questions

1. Since \hat{z} is not part of the training dataset, how do we know that its effect is unlearned appropriately from the model with the elastic weight consolidation loss ? 2. How sensitive is \hat{z} to subtle changes in the caption ? 3. Have you considered comparing your work against DataInf: https://arxiv.org/abs/2310.00902 ?

Rating

6

Confidence

3

Soundness

3

Presentation

3

Contribution

3

Limitations

The paper discusses the limitations of the work.

Reviewer 8wwb6/10 · confidence 2/52024-07-16

Summary

This paper proposes a novel method for data attribution in text-to-image diffusion models. The key idea is to unlearn a synthesized image by optimizing the model to increase its loss on that image, while using elastic weight consolidation to avoid catastrophic forgetting. The authors then identify influential training images by measuring which ones have the largest increase in loss after this unlearning process. The method is evaluated through rigorous counterfactual experiments on MSCOCO, where models are retrained after removing the identified influential images. It outperforms baselines like influence functions and feature matching approaches. The method is also tested on a benchmark for attributing customized models.

Strengths

1. The paper addresses the issue of data attribution, which is important for understanding the behavior of generative models and the contribution of training data. 2. The paper proposes an innovative and effective approach to data attribution that outperforms existing methods on counterfactual evaluations. 3. The authors conduct extensive experiments, including computationally intensive retraining, to thoroughly validate the effectiveness of the proposed method.

Weaknesses

1. The evaluations are limited to medium-scale datasets like MSCOCO. It's unclear how well the approach would scale to the massive datasets, such as LAION, which is used to train state-of-the-art text-to-image models. 2. The method is computationally expensive, requiring many forward passes over the training set to estimate losses. This may limit scalability to large datasets.

Questions

1. How does the performance of the unlearning method vary with different sizes and types of datasets? 2. How does the unlearning method perform in terms of efficiency compared with other methods?

Rating

6

Confidence

2

Soundness

3

Presentation

3

Contribution

3

Limitations

Yes

Reviewer ZhUS7/10 · confidence 3/52024-07-22

Summary

Authors propose a method for identifying training images that need to be removed from the training set of a generative model to prevent a single specific “bad” (undesired) output from occurring in its output. Authors propose to directly unlearn the synthetized “bad” image, evaluate how this unlearning changes the training loss across training examples, and retrain the model omitting highly effected training examples - instead of evaluating how including each training example affects the “bad” output. Authors found that the resulting approach worked best when combined with prior techniques for preventing catastrophic forgetting in classifier uncleaning (e.g. optimizing only a set of weights and performing unlearning steps using Newton-Raphson iterations on the unlearned example using the original Fisher matrix). Authors show that retraining generators on pruned training sets indeed prevents the retrained generator from generating the undesired image - more specifically, authors show that the loss on the undesired output increases (more so compared to baseline methods) and that rerunning the retrained generator with the same random seed and text conditioning does not produce the “bad” example indeed.

Strengths

While at first the result might seem trivial (since authors both optimize and measure the same metric - “the model loss”) the finding is actually in fact very much non-trivial since authors retrain the model while pruning images that were affected most by unlearning the problematic output, so it might have so happened that removing these images did not prevent the output from occurring - so reported findings are indeed both interesting and valuable for the community (potentially beyond unlearning work). Authors provide a random baseline and all provided results are significantly different from random (confidence intervals seem unnecessary). Authors provide qualitative evaluation (Fig 3) suggesting that the desired effect is indeed observed in practice and not only in metrics. Overall, the paper is well written and does a very good job of motivating both the problem and the proposed solution. Authors provide an extensive literature review that manages to be helpful even to a person without prior experience with unlearning (me).

Weaknesses

While I enjoyed reading the first half of the paper, I have serious concerns regarding the quality of the results section. While authors provide many qualitative results in the main paper, the burden of making sense of quantitative results (e.g. comparing them to ablations) is not only entirely on the reader, but is made worse by complete lack of any raw numbers for cross-comparison (e.g. no tables - only barplots w\o numbers), and putting the vast majority of quantitative results into supplementary without providing that much more context there. For example, Figure 4 (top) - colors are very similar and lines are very close, making results not really legible. Given that performance is evaluated only at three data points, I see no reason why this can not be a table - that would aid both reproducibility and comprehension (e.g. cross-comparison). I also do not quite understand the point of reporting the "equivalent to X random points” metric - it is never even explicitly discussed in the main paper. All results for MSE and CLIP metric and all Ablations (also no tables, only bar plots without numbers) are in the supplementary, but even there not much additional context for interpreting these results is provided. Some claims are discussed and are just never qualitatively verified - e.g. authors claim that the proposed technique is specifically designed to not lead to catastrophic forgetting, but I could not find any evidence that would support that. Moreover, from Figure 3 it might seem like it not only forgot how to generate this specific bus, but might have forgotten how to generate all buses - I am not entirely sure if that is the intended behavior. I would appreciate a qualitative confirmation that the model did not just “forget everything”. I have minor concerns regarding the delta G metric. While Georgiev et al. [12] supposedly already showed in an accepted (workshop) submission that two models trained independently on the same or similar datasets generate similar images when primed with the same noise input - justifying the existence of metric (3) - I’d appreciate some results confirming that this is indeed the case for these models in these experiments. For example, if authors showed at least qualitatively that removing images most influential for generating a particular “bad” bedroom from the training set does not indeed in any way affect the generation of unrelated images (e.g. an image of mountains), that would help. Or, even better, by performing textual inversion on the “undesired” input after removing images that affect it the most, to show that it can no longer be represented by the model. Otherwise, there is no way to tell if a particular “bad generation” just “migrated” elsewhere. Another minor concern: since authors performed multiple gradient update steps, I wonder whether other methods can also be used with multiple update steps.

Questions

In addition to the weaknesses listed above, here are some questions/concerns I had while reading the paper: Figure 1 - it is not said explicitly, but after several attempts, I figured out that the “bad” output that we want to remove in this figure is “bedrooms”, so it makes sense that the method assigned high loss (bottom orange bars) and “picked” bedroom images, the fact that only a single image (mountain) either “picked by the method” or “remains in the training set” is confusing; it might be worth to keep only two bedrooms and add a couple more images with low loss and somehow highlighting that “bedrooms are picked for removal”, not “only a single mountain is picked for keeping”. L130 vs L104 - it is not clear why you need to introduce \tilda \theta if you already have \theta, especially given that in (2) RHS appears to not depend on \theta (w\o \tilda) and L157 also uses \theta (w\o \tilda), so the use of \tilda there is somewhat inconsistent and confusing; looking at L483-499 I think I understand now that the implied difference between the two is “optimal” vs “updated”, but i think this distinction is somewhat confusing in the main paper L132 - \epsilon is never introduced at this point, so this is a bit confusing L137-139 - trained from the same random noise or generated from the same random noise? worth rewording L168-172 - these lines appear to not follow from the previous paragraphs introducing the notation for the attribution algorithm since it already provides per-sample scores (and therefore does not require 2^K evaluations); or not, and each tau(\hat z, z_i) is assumed to potentially involve iterating over all subsets of the training set (w\ and w\o z_i) - if so, please elaborate Figure 2 - “Notably, our method better matches the poses of the buses (considering random flips during training) and the poses and enumeration of skiers.” - I would not say that it is very apparent from the figure.

Rating

7

Confidence

3

Soundness

3

Presentation

2

Contribution

3

Limitations

Authors do mention some limitations of their work.

Reviewer UBLR2024-08-07

Response to the authors

Thank you authors for the detailed response. I have a question while looking at Table 1 in the PDF. Is this table for 1 generated image or is this an aggregate for multiple generated images ? How is the domain of those generated images identified ? How do we ensure that our evaluation of generated images is representative ? How does the generation change if we say: "bus", "white bus" vs "big white bus" for K-leave out test in figure 1? In Figure 1 are `Related images` the ones that are being removed during K-leave out experiment ? If not then it is not clear which images are left out for K-leave out experiment. Is it possible to include textual captions for generation into the visual results ?

Authorsrebuttal2024-08-11

Thanks for your quick response. - Table 1 is consistent with the practice in the paper, as described in L236. The results are aggregated by 110 generated images. - The generated images' prompts are from the MSCOCO validation set. This ensures that generated images are representative of the training set (also MSCOCO). - The generated images don’t change much when we change “bus” in the caption to “white bus” or “big white bus”. - No, only the attributed training images from the target are removed (see top row of Fig. 2, main paper). The “related images” are *generated* images, from captions similar to the target prompt. As seen in the figure, the leave-K-out model successfully removes the target image while preserving the related images. Please see general response G2 for more detail. - Yes, we will include captions when adding to our revision, similar to Figs 2&3 in the main paper.

Reviewer UBLR2024-08-13

Response to authors

Thank you for the clarification! If related images are the ones generated based on the captions similar to the target it is important to show those captions. Also, how is "similar" measured here. Since the authors mostly answered my questions I will increase the score but I encourage the authors to incorporate the clarifications and suggestions made in the rebuttal.

Authorsrebuttal2024-08-13

We generate related images using the most similar 100 captions retrieved from the MSCOCO val set using CLIP's text encoder. Also, thanks for your comments and updated scores. Yes, we will incorporate the clarifications and suggestions.

Reviewer 8wwb2024-08-09

Thank you for your response. The authors have addressed my concerns. Based on their rebuttal, I have decided to upgrade my evaluation to the weak accept.

Authorsrebuttal2024-08-11

Thanks for increasing the rating. We will incorporate the feedback in our revision.

Reviewer ZhUS2024-08-09

rebuttal

I appreciate the effort to explain the original motivation (for metrics used, etc.) and I agree with it. I also find Table 1 in the rebuttal pdf much more convincing (and thank you for error bars). Having read the rebuttal message, pdf, and rebuttals for other reviews, I think the final paper would be greatly improved by the addition of results and formatting changes (format results as tables) proposed during review. I think Figure 1 in the rebuttal also mostly addresses my concerns re G delta metric (shows that other images are indeed reconstructed almost perfectly confirming claims of Georgiev et al). For completeness, I'd appreciate Related/Other MSE/CLIP measurement for other baselines (if images are generated and stored, this should not be difficult)? And (later) I would encourage authors to put in more work into Textual Inversion experiments - people report much better reconstruction quality than what was provided in the rebuttal Fig 3. With the addition of (many) quantitative and qualitative evaluations added in the rebuttal, this submission shapes into a convincing well motivated work. My only concern is that with that many changes, a significant rewrite of the second half of the paper will take place - and that will not be peer-reviewed. Given that the authors did a good job with both 1) the first half of the original submission and 2) provided convincing results in this rebuttal - I tend to believe that they will be able to rewrite the second half of the paper incorporating all the feedback and experiments provided above. In the light of the previous paragraph, I increased my final rating to Accept, but I encourage authors to do a thorough reword of the second half of the paper.

Authorsrebuttal2024-08-11

Thanks for increasing the rating. We will incorporate the feedback in our revision and reword/reorganize the second half of the paper.

Program Chairsdecision2024-09-25

Decision

Accept (poster)

© 2026 NYSGPT2525 LLC