Pith. sign in

REVIEW 4 major objections 5 minor 52 references

Where's the liability in the Generative Era? Recovery-based Black-Box Detection of AI-Generated Content

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A black-box detector can tell AI-generated images from real photos by masking part of an image and measuring how well the suspected generative model repairs it, reaching 86.61% mean average precision across eight diffusion-model variants…

desk verdict A practical LoRA-aligned inpainting detector for black-box AI images, but the main result is not trustworthy until the alignment/test split is clarified and DIRE is added as a baseline. read the letter →

arxiv 2505.01008 v2 pith:PI2H6TLP submitted 2025-05-02 cs.LG

Haoyue Bai , Yiyou Sun , Wei Cheng , Haifeng Chen This is my paper · ORCID
classification cs.LG
keywords AI-generatedimagedetectionblack-boxdiffusionmodelscorrupt-and-recoverLoRAfine-tuninglikelihoodgapPSNRrecoveryscoreDALL-E3benchmark
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that AI-generated images can be reliably detected without any access to model weights or large training datasets: you only need the suspect model's API. The central trick is a corrupt-and-recover test: mask part of an image, ask the model to fill in the missing region, and measure how faithfully it reconstructs it. Because a generative model should recognize its own outputs, its repairs to its own images are closer to the originals than its repairs to real photographs, so the recovery score separates fake from real. For closed APIs that only generate from scratch, the paper fine-tunes an open surrogate model with LoRA on fewer than 1,000 samples from the target API to align the two distributions. The result is 86.61% mean average precision across eight diffusion-model variants, beating supervised baselines by 4.31%.

What carries the argument

The load-bearing object is the corrupt-and-recover discrepancy score: for an image x, a binary mask m splits it into known pixels X=(1−m)⊙x and unknown pixels Y0=m⊙x; the model samples K reconstructions Y′∼G(·|X), and the PSNR (Peak Signal-to-Noise Ratio) between Y0 and the recovered region is the detection statistic. It is paired with the Likelihood-Gap Hypothesis, which asserts that E_{Y∼G(·|X)}[log p(Y|X)]−E_{Y∼H(·|X)}[log p(Y|X)]>Δ, the formal reason a model should repair its own outputs better. The third component is distribution alignment: LoRA fine-tuning of an open surrogate on fewer than 1,000 target-API samples synchronizes the surrogate's distribution with the closed target model, letting the recovery test work when the API only generates from scratch. The theory also fixes the number of recovery samples at K=Ω(σ log(1/δ)/$Δ^{2}$) via a Hoeffding and sub-Gaussian argument.

What would settle it

Take the paper's paired DALL-E 3 benchmark, mask half of each image, compute PSNR recovery scores with a surrogate aligned to DALL-E 3, and estimate the likelihood-gap Δ between fake and real images; if the AUROC of the recovery score approaches 0.5 (or Δ is statistically indistinguishable from zero) once image resolution and photorealism are high enough, the central claim is falsified for that model.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a generative model should more easily recover its own generated images when corrupted than it would with real images. The detector masks a region, samples K recoveries from the model conditioned on the unmasked pixels, and scores the discrepancy with PSNR; a high recovery score marks the image as machine-generated. For black-box APIs that do not accept masked inputs, the paper aligns an open-source surrogate (Stable Diffusion) to the target distribution by LoRA fine-tuning on a small set of API samples, then runs the same recovery test with the surrogate. The formal statement is the Likelihood-Gap Hypothesis: conditioned on the known pixels, the expected log-likelihood of the machine generation process exceeds that of the real-image process by a positive gap $\Delta$, and the paper argues this gap makes reliable detection possible even when the total-variation distance is small. Empirically the framework reports 86.61% mean average precision across eight diffusion-model variants (Guided, LDM, Glide, DALL-E, DALL-E 3), including a new paired DALL-E 3 benchmark on which human accuracy is about 72%.

Load-bearing premise

The entire detector depends on the assumption that a machine generation process assigns measurably higher conditional likelihood to its own masked images than the real-photo process does; for a photorealistic model such as DALL-E 3, if this likelihood gap shrinks to zero, the recovery score no longer separates fake from real.

Editorial extensions

If this is right

  • AI-content detection no longer requires white-box access: any model reachable through an API can be tested by its own repair behavior, so detectors scale to closed commercial generators.
  • A single open surrogate (Stable Diffusion) aligned with a small API sample set detects images from many unseen diffusion architectures, so per-model retraining is unnecessary.
  • The choice of discrepancy metric matters materially: PSNR outperforms SSIM, L1, and L2 by large margins, so future detectors should treat the scoring function as a first-class design choice.
  • In the white-box setting, using the target model itself for recovery gives the strongest separation (e.g., FPR 10.80% and AUROC 97.18% on Guided Diffusion), so the same framework covers both settings.
  • Detection accuracy falls on photorealistic DALL-E 3 images (75.98% AP) even though humans only reach about 72% accuracy, indicating the recovery signal survives but weakens as generation quality improves.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the Likelihood-Gap Hypothesis generalizes, the same mask-and-repair principle could be applied to video, audio, and text by substituting a domain-appropriate recovery model and scoring function; the paper only demonstrates images.
  • The paper's own DALL-E 3 numbers suggest the gap Δ shrinks as generators approach the real-image manifold; a testable prediction is that for future photorealistic models, the PSNR separation will narrow and LoRA alignment will become the dominant factor.
  • The recovery score could be repurposed as a model-attribution or membership test: among several candidate APIs, the model that repairs an image best is likely its source, which would extend binary fake/real detection to source identification.
  • An adversary aware of the detector could potentially craft images whose masked regions are hard to repair, so robustness of the recovery score under adversarial generation is an open question the paper does not address.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The manuscript proposes a black-box detector for AI-generated images based on a corrupt-and-recover principle: a surrogate generative model, aligned to the target model via LoRA fine-tuning on a small set of API-generated images, reconstructs masked regions, and a PSNR-based score separates real from generated images. Experiments report 86.61% mean average precision across eight diffusion-model variants using a Stable Diffusion surrogate, with less than 1,000 alignment samples and under two GPU hours. The paper also introduces a paired DALL-E 3 benchmark and a human-evaluation study, and releases code.

Significance. If the empirical claims hold with a clean evaluation protocol, the method is practically valuable: it removes the need for model weights and large real-image corpora, and it offers a resource-efficient black-box detection route. The idea of aligning a surrogate by parameter-efficient fine-tuning is sound and the reported gains over the included baselines are substantial. The paper also ships code and a new paired benchmark, which are useful community resources. However, the central empirical claim currently rests on an unstated train/test split for the alignment set, and the most relevant baseline, DIRE, is missing; both issues must be resolved before the stated performance numbers can be taken at face value.

major comments (4)
  1. [§3.2 and §4.1, Table 1] The manuscript never states whether the alignment dataset S, collected from the target-model API in Section 3.2, is disjoint from the fake test images used to compute the detection scores in Table 1. This is load-bearing: if S contains or overlaps the test generated images, the LoRA-fine-tuned surrogate is evaluated on images it was fine-tuned on, so the reported 86.61 mAP and the Table 2 improvement from 86.74 to 92.97 AP after fine-tuning could reflect memorization rather than generalization. Please specify the exact split (e.g., generate a fresh set of target-model images for testing after collecting S) and rerun Table 1 and Table 2 on a truly held-out set of target-generated images, reporting the numbers for both the overlapped and disjoint settings.
  2. [§2.2 and §4.2, Table 1] DIRE [43] is cited in Related Work as observing that diffusion-generated images can be approximately reconstructed by a diffusion model while real images cannot, which is the same recovery-based principle used here. DIRE is therefore the most directly relevant baseline, yet it is not included in Table 1 or anywhere in the experiments. Without a DIRE comparison, the claim of outperforming baseline methods by 4.31% mAP is not fully supported. Please add DIRE to Table 1 and to the new DALL-E 3 benchmark, using the same evaluation protocol.
  3. [§3.3 and Appendix A.1] The Likelihood-Gap Hypothesis assumes a positive gap Delta between the expected log-likelihood of the machine generation process G and the real process H, but Delta is not derived and H is not defined as a concrete process for real photographs. Appendix A.1 then uses this hypothesis to conclude that detection is always possible, but the conclusion is conditional on the assumed gap, not a guarantee. Please clarify the status of the hypothesis: state explicitly that it is an assumption, provide a definition of H, and either prove or empirically test the existence of the gap for the models considered, or weaken the theoretical claims accordingly.
  4. [§4.2, DALL-E 3 benchmark] The paper introduces a new paired DALL-E 3 benchmark, but no detector results are reported on it. The benchmark is only used for the human-preference study in Section 4.4. Since the DALL-E 3 setting is described as especially challenging and the paper claims a practical detector, please report the proposed method's AP/AUROC/FPR95 on this benchmark and compare with at least the strongest baselines from Table 1, or explicitly state that the benchmark is released for future work only.
minor comments (5)
  1. [§3, Preliminary on Diffusion Models] The word 'numerious' should be 'numerous'.
  2. [§4.3, Visualization] The word 'obervations' should be 'observations' in the visualization paragraph.
  3. [§4.1, Metrics] The threshold definition 'selected based on the fake data when 95% of fake test data points are declared as fake' is ambiguous; please clarify whether this is FPR at 95% recall on the fake class or another convention, and state the threshold-selection protocol used for the other metrics.
  4. [Table 1 and Table 2] The manuscript alternates between 'AP' and 'mAP' for the same column; please use one consistent notation and define it in the caption.
  5. [§4.2, New dataset for DALL-E 3] The paper does not specify how many image pairs are in the new DALL-E 3 benchmark or how the LLaVA-generated captions were filtered; please add dataset statistics and a brief filtering description.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the recovery hypothesis is explicit and empirically supported, and LoRA alignment is a transfer step rather than a fitted prediction.

full rationale

The paper's claimed derivation chain is: (1) a generative model recovers its own outputs better than real images (an explicit hypothesis); (2) this recovery gap is measured by a discrepancy score; (3) for black-box APIs, a surrogate is LoRA-tuned on target-model images; (4) the resulting score is evaluated for real-vs-fake discrimination. Step (1) is labeled a hypothesis, not derived from the detection target, and it is supported by Figure 1 and the ablations in Tables 2 and 3. Step (3) is a genuine transfer step: the surrogate is fine-tuned on target images and then tested on the real-vs-fake classification task, so the target quantity is the gap between classes, not the fine-tuning loss itself. The single self-citation [46] is used only to motivate the intuition and is not load-bearing for the empirical claims, which are compared against external baselines. The theoretical appendix explicitly calls the gap a hypothesis and derives only conditional guarantees, such as the sample complexity of K; no equation in the paper reduces to its own input by construction. The one serious concern is experimental hygiene: the paper does not state whether the alignment set S is disjoint from the fake test images in Table 1. If S overlapped the test fake set, the reported AP could be inflated by memorization, but the paper does not define the test set as a subset of S, so this remains a reproducibility risk rather than an in-text circular reduction. Under the hard rule requiring quoted evidence of a specific construction, no circular step is established.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

Everything beyond standard diffusion background rests on the assumed Likelihood-Gap Hypothesis, the sub-Gaussian distance assumption for sample complexity, and the pairing of LAION with the real-image process H. The LoRA alignment step introduces a practical dependence on an unstated alignment dataset size and undisclosed test overlap.

free parameters (3)
  • Decision threshold tau = not reported
    Equation (1) classifies by comparing the discrepancy score to tau; FPR95 is defined by selecting a threshold from fake test data, so the operational detector depends on this fitted threshold.
  • Recovery sample count K = not reported
    Section 3.3 and Appendix A.2 state K should scale as Omega(sigma log(1/delta)/Delta^2), but the paper does not report sigma, Delta, or the actual K used in the experiments.
  • Alignment dataset size N = claimed under 1,000
    Section 3.2 collects N images from the target API for LoRA fine-tuning; the exact N, how it varies per model, and whether these images are disjoint from evaluation images are not given, which matters for both cost claims and leakage.
assumptions (4)
  • domain assumption Likelihood-Gap Hypothesis: E_Y~G[log p(Y|X)] - E_Y~H[log p(Y|X)] > Delta > 0 for masked inputs.
    Assumed in Section 3.3 and used in Appendix A.1 to rebut impossibility results; no derivation or empirical estimate of the gap is provided.
  • domain assumption Assumption 1: the distances D(Y0,Y') and D(~Y0,Y') are sigma-sub-Gaussian with a positive distance gap Delta.
    Stated in Appendix A.2 to justify the sample complexity bound for K; the values of sigma and Delta are unknown and not estimated from data.
  • domain assumption LAION images represent the real-world distribution H for all tested domains.
    Section 4.1 uses LAION as the real class; the theoretical statements about real images implicitly identify H with this dataset.
  • domain assumption A LoRA fine-tuned open-source surrogate faithfully approximates the target model's distribution for masked recovery.
    Section 3.2 assumes distribution alignment can be achieved with fewer than 1,000 samples; Table 2 shows an improvement on one target model but does not establish fidelity for all targets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Where's the liability in the Generative Era? Recovery-based Black-Box Detection of AI-Generated Content." pith.science (2026). https://pith.science/paper/PI2H6TLP

@misc{pith2026250501008,
  author       = {Pith},
  title        = {Pith review of: Where's the liability in the Generative Era? Recovery-based Black-Box Detection of AI-Generated Content},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PI2H6TLP}},
  note         = {Machine review of arXiv:2505.01008}
}
read the original abstract

The recent proliferation of photorealistic images created by generative models has sparked both excitement and concern, as these images are increasingly indistinguishable from real ones to the human eye. While offering new creative and commercial possibilities, the potential for misuse, such as in misinformation and fraud, highlights the need for effective detection methods. Current detection approaches often rely on access to model weights or require extensive collections of real image datasets, limiting their scalability and practical application in real world scenarios. In this work, we introduce a novel black box detection framework that requires only API access, sidestepping the need for model weights or large auxiliary datasets. Our approach leverages a corrupt and recover strategy: by masking part of an image and assessing the model ability to reconstruct it, we measure the likelihood that the image was generated by the model itself. For black-box models that do not support masked image inputs, we incorporate a cost efficient surrogate model trained to align with the target model distribution, enhancing detection capability. Our framework demonstrates strong performance, outperforming baseline methods by 4.31% in mean average precision across eight diffusion model variant datasets.

Figures

Figures reproduced from arXiv: 2505.01008 by the authors.

Figure 1
Figure 1. Visual quality comparison of surrogate model recovery [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An overview of our proposed black-box content detection framework. Given a candidate input image, we aim to determine [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Evaluation on differen scores and evaluation metrics. The PSNR scores here show better detection accuracy in terms of FPR95, [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Different evaluation scores demonstrate significant dif [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 6
Figure 6. Figure 6: Examples of hard cases for distinction in human evaluations. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Example of different masks. photorealistic image generation and editing. It employs classifier-free guidance to enhance image quality while maintaining fidelity to text prompts. LDM [33] apply diffusion processes in the latent space of pretrained autoencoders rather th…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 26 canonical work pages

  1. [43]

    Dire for diffusion-generated image detection

    Zhendong Wang, Jianmin Bao, Wengang Zhou, Weilun Wang, Hezhen Hu, Hong Chen, and Houqiang Li. Dire for diffusion-generated image detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 22445–22455, 2023. 3

    2023

  2. [1]

    Feed two birds with one scone: Exploiting wild data for both out-of-distribution gen- eralization and detection

    Haoyue Bai, Gregory Canal, Xuefeng Du, Jeongyeol Kwon, Robert D Nowak, and Yixuan Li. Feed two birds with one scone: Exploiting wild data for both out-of-distribution gen- eralization and detection. In International Conference on Machine Learning, pages 1454–1471. PMLR, 2023. 2

    2023

  3. [2]

    Synthbuster: Towards detection of diffu- sion model generated images

    Quentin Bammey. Synthbuster: Towards detection of diffu- sion model generated images. IEEE Open Journal of Signal Processing, 2023. 2

    work page 2023

  4. [3]

    https://openai.com/dall-e-3 ,

    James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, Wesam Manassra, Prafulla Dhariwal, Casey Chu, and Yunxin Jiao. https://openai.com/dall-e-3 ,

    work page

  5. [4]

    Improving image generation with better captions

    James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, et al. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3):8, 2023. 1, 2

    2023

  6. [5]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale gan training for high fidelity natural image synthesis. arXiv preprint arXiv:1809.11096, 2018. 2

    arXiv 2018

  7. [6]

    What makes fake images detectable? understanding proper- ties that generalize

    Lucy Chai, David Bau, Ser-Nam Lim, and Phillip Isola. What makes fake images detectable? understanding proper- ties that generalize. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXVI 16, pages 103–120. Springer, 2020. 1, 2, 5, 6

    work page 2020

  8. [7]

    Muse: Text-to-image generation via masked generative transform- ers

    Huiwen Chang, Han Zhang, Jarred Barber, AJ Maschinot, Jose Lezama, Lu Jiang, Ming-Hsuan Yang, Kevin Mur- phy, William T Freeman, Michael Rubinstein, et al. Muse: Text-to-image generation via masked generative transform- ers. arXiv preprint arXiv:2301.00704, 2023. 2

    arXiv 2023

Show all 52 references
  1. [8]

    Stargan: Unified genera- tive adversarial networks for multi-domain image-to-image translation

    Yunjey Choi, Minje Choi, Munyoung Kim, Jung-Woo Ha, Sunghun Kim, and Jaegul Choo. Stargan: Unified genera- tive adversarial networks for multi-domain image-to-image translation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8789–8797,

  2. [9]

    On the detection of synthetic images generated by diffusion mod- els

    Riccardo Corvi, Davide Cozzolino, Giada Zingarini, Gio- vanni Poggi, Koki Nagano, and Luisa Verdoliva. On the detection of synthetic images generated by diffusion mod- els. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), p...

    2023

  3. [10]

    Forensictrans- fer: Weakly-supervised domain adaptation for forgery detec- tion

    Davide Cozzolino, Justus Thies, Andreas R ¨ossler, Christian Riess, Matthias Nießner, and Luisa Verdoliva. Forensictrans- fer: Weakly-supervised domain adaptation for forgery detec- tion. arXiv preprint arXiv:1812.02510, 2018. 3

    2018 arXiv

  4. [11]

    Diffusion models beat gans on image synthesis, 2021

    Prafulla Dhariwal and Alex Nichol. Diffusion models beat gans on image synthesis, 2021. 5

    2021

  5. [12]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural informa- tion processing systems, 34:8780–8794, 2021. 2, 12

    2021

  6. [13]

    Lighting (in) consistency of paint by text

    Hany Farid. Lighting (in) consistency of paint by text. arXiv preprint arXiv:2207.13744, 2022. 3

    2022 arXiv

  7. [14]

    Perspective (in) consistency of paint by text

    Hany Farid. Perspective (in) consistency of paint by text. arXiv preprint arXiv:2206.14617, 2022. 3

    2022 arXiv

  8. [15]

    Are gan generated im- ages easy to detect? a critical analysis of the state-of-the-art

    Diego Gragnaniello, Davide Cozzolino, Francesco Marra, Giovanni Poggi, and Luisa Verdoliva. Are gan generated im- ages easy to detect? a critical analysis of the state-of-the-art. In 2021 IEEE international conference on multimedia and expo (ICME), pages 1–6. IEEE, 2021. 3

    2021

  9. [16]

    Denoising diffusion models for out-of-distribution detection

    Mark S Graham, Walter HL Pinaya, Petru-Daniel Tudosiu, Parashkev Nachev, Sebastien Ourselin, and Jorge Cardoso. Denoising diffusion models for out-of-distribution detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2948–2957, ...

    2023

  10. [17]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 5

    2016

  11. [18]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 3, 4

    2020

  12. [19]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685, 2021. 4

    2021 arXiv

  13. [20]

    Fake trump arrest photos: How to spot an ai-generated image

    Kayleen Devlin and Joshua Cheetham. Fake trump arrest photos: How to spot an ai-generated image. BBC News ,

  14. [21]

    Single-model attribution of generative models through final-layer inversion

    Mike Laszkiewicz, Jonas Ricker, Johannes Lederer, and Asja Fischer. Single-model attribution of generative models through final-layer inversion. arXiv preprint arXiv:2306.06210, 2023. 1

    2023 arXiv

  15. [22]

    Asymptotic methods in statistical decision theory

    Lucien Le Cam. Asymptotic methods in statistical decision theory. Springer Science & Business Media, 2012. 11

    2012

  16. [23]

    Detecting gan-generated images by orthogo- nal training of multiple cnns

    Sara Mandelli, Nicol `o Bonettini, Paolo Bestagini, and Ste- fano Tubaro. Detecting gan-generated images by orthogo- nal training of multiple cnns. In 2022 IEEE International Conference on Image Processing (ICIP), pages 3091–3095. IEEE, 2022. 3

    2022

  17. [24]

    Do gans leave artificial fingerprints? In 2019 IEEE conference on multimedia information process- ing and retrieval (MIPR), pages 506–511

    Francesco Marra, Diego Gragnaniello, Luisa Verdoliva, and Giovanni Poggi. Do gans leave artificial fingerprints? In 2019 IEEE conference on multimedia information process- ing and retrieval (MIPR), pages 506–511. IEEE, 2019. 2

    2019

  18. [25]

    Ex- ploiting visual artifacts to expose deepfakes and face manip- ulations

    Falko Matern, Christian Riess, and Marc Stamminger. Ex- ploiting visual artifacts to expose deepfakes and face manip- ulations. In 2019 IEEE Winter Applications of Computer Vision Workshops (WACVW), pages 83–92. IEEE, 2019. 2

    2019

  19. [26]

    Detecting gan- generated imagery using color cues

    Scott McCloskey and Michael Albright. Detecting gan- generated imagery using color cues. arXiv preprint arXiv:1812.08247, 2018. 2

    2018 arXiv

  20. [27]

    Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021. 2, 3, 5, 12

    2021 arXiv

  21. [28]

    Towards uni- versal fake image detectors that generalize across genera- tive models

    Utkarsh Ojha, Yuheng Li, and Yong Jae Lee. Towards uni- versal fake image detectors that generalize across genera- tive models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24480– 24489, 2023. 2, 3, 5

    2023

  22. [29]

    Im- age transformer

    Niki Parmar, Ashish Vaswani, Jakob Uszkoreit, Lukasz Kaiser, Noam Shazeer, Alexander Ku, and Dustin Tran. Im- age transformer. In International conference on machine learning, pages 4055–4064. PMLR, 2018. 2

    2018

  23. [30]

    Zero-shot text-to-image generation

    Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea V oss, Alec Radford, Mark Chen, and Ilya Sutskever. Zero-shot text-to-image generation. In International confer- ence on machine learning, pages 8821–8831. Pmlr, 2021. 1, 5, 12

    2021

  24. [31]

    Hierarchical text-conditional image gener- ation with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image gener- ation with clip latents. arXiv preprint arXiv:2204.06125, 1 (2):3, 2022. 2, 3

    2022 arXiv

  25. [32]

    Towards the detection of diffusion model deepfakes

    Jonas Ricker, Simon Damm, Thorsten Holz, and Asja Fis- cher. Towards the detection of diffusion model deepfakes. arXiv preprint arXiv:2210.14571, 2022. 3

    2022 arXiv

  26. [33]

    High-resolution image syn- thesis with latent diffusion models, 2021

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High-resolution image syn- thesis with latent diffusion models, 2021. 12

    2021

  27. [34]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10684–10695, 2022. 4

    2022

  28. [35]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. 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. 1, 2, 3, 5

    2022

  29. [36]

    Can ai-generated text be reliably detected? arXiv preprint arXiv:2303.11156,

    Vinu Sankar Sadasivan, Aounon Kumar, Sriram Balasubra- manian, Wenxiao Wang, and Soheil Feizi. Can ai-generated text be reliably detected? arXiv preprint arXiv:2303.11156,

    arXiv

  30. [37]

    Laion-400m: Open dataset of clip-filtered 400 million image-text pairs

    Christoph Schuhmann, Richard Vencu, Romain Beaumont, Robert Kaczmarczyk, Clayton Mullis, Aarush Katta, Theo Coombes, Jenia Jitsev, and Aran Komatsuzaki. Laion-400m: Open dataset of clip-filtered 400 million image-text pairs. arXiv preprint arXiv:2111.02114, 2021. 5

    2021 arXiv

  31. [38]

    Laion-5b: An open large-scale dataset for training next generation image-text models

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Worts- man, et al. Laion-5b: An open large-scale dataset for training next generation image-text models. Advances in Neural In- f...

    2022

  32. [39]

    De-fake: Detection and attribution of fake images generated by text- to-image generation models

    Zeyang Sha, Zheng Li, Ning Yu, and Yang Zhang. De-fake: Detection and attribution of fake images generated by text- to-image generation models. In Proceedings of the 2023 ACM SIGSAC Conference on Computer and Communica- tions Security, pages 3418–3432, 2023. 3

    2023

  33. [40]

    Gemini 1.5: Unlocking multimodal understanding across millions of tokens of con- text

    Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, et al. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of con- text. arXiv preprint arXiv:2403.05530, 2024. 2

    2024 arXiv

  34. [41]

    Uk engineering giant arup falls victim to deepfake scam in hong kong

    The Guardian. Uk engineering giant arup falls victim to deepfake scam in hong kong. The Guardian, 2024. 1

    2024

  35. [42]

    Cnn-generated images are surprisingly easy to spot

    Sheng-Yu Wang, Oliver Wang, Richard Zhang, Andrew Owens, and Alexei A Efros. Cnn-generated images are surprisingly easy to spot... for now. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8695–8704, 2020. 2, 3, 5, 6

    2020

  36. [44]

    Where did i come from? origin attribution of ai-generated images

    Zhenting Wang, Chen Chen, Yi Zeng, Lingjuan Lyu, and Shiqing Ma. Where did i come from? origin attribution of ai-generated images. Advances in neural information pro- cessing systems, 36, 2024. 1, 3

    2024

  37. [45]

    Lecture notes for stat 705: Advanced data analysis, 2023

    Larry Wasserman. Lecture notes for stat 705: Advanced data analysis, 2023. 11

    2023

  38. [46]

    Dna-gpt: Divergent n-gram analysis for training-free detection of gpt-generated text

    Xianjun Yang, Wei Cheng, Yue Wu, Linda Ruth Petzold, William Yang Wang, and Haifeng Chen. Dna-gpt: Divergent n-gram analysis for training-free detection of gpt-generated text. In The Twelfth International Conference on Learning Representations. 2

  39. [47]

    Attributing fake images to gans: Learning and analyzing gan fingerprints

    Ning Yu, Larry S Davis, and Mario Fritz. Attributing fake images to gans: Learning and analyzing gan fingerprints. In Proceedings of the IEEE/CVF international conference on computer vision, pages 7556–7566, 2019. 2

    2019

  40. [48]

    Styleswin: Transformer-based gan for high-resolution image genera- tion

    Bowen Zhang, Shuyang Gu, Bo Zhang, Jianmin Bao, Dong Chen, Fang Wen, Yong Wang, and Baining Guo. Styleswin: Transformer-based gan for high-resolution image genera- tion. In Proceedings of the IEEE/CVF conference on com- puter vision and pattern recognition , pages 11304–11314,

  41. [49]

    Detecting and simulating artifacts in gan fake images

    Xu Zhang, Svebor Karaman, and Shih-Fu Chang. Detecting and simulating artifacts in gan fake images. In2019 IEEE in- ternational workshop on information forensics and security (WIFS), pages 1–6. IEEE, 2019. 2, 3, 5, 6

    2019

  42. [50]

    Unpaired image-to-image translation using cycle- consistent adversarial networks

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision , pages 2223– 2232, 2017. 2

    2017

  43. [51]

    Genimage: A million-scale benchmark for de- tecting ai-generated image

    Mingjian Zhu, Hanting Chen, Qiangyu Yan, Xudong Huang, Guanyu Lin, Wei Li, Zhijun Tu, Hailin Hu, Jie Hu, and Yunhe Wang. Genimage: A million-scale benchmark for de- tecting ai-generated image. Advances in Neural Information Processing Systems, 36, 2024. 2

    2024

  44. [52]

    Conditional text image generation with diffu- sion models

    Yuanzhi Zhu, Zhaohai Li, Tianwei Wang, Mengchao He, and Cong Yao. Conditional text image generation with diffu- sion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14235– 14245, 2023. 2 A. Theoretical Analysis A.1. Is it alw...

    2023

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.