SFE-Net: Harnessing Biological Principles of Differential Gene Expression for Improved Feature Selection in Deep Learning Networks

Haojun Fei; Jianjun Yin; Yaoxuan Wang; Yuanzhong Zheng; Yuqi Li

REVIEW 4 major objections 6 minor 26 references

SFE-Net: Harnessing Biological Principles of Differential Gene Expression for Improved Feature Selection in Deep Learning Networks

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read SFE-Net claims a dynamic feature-selection module lifts deepfake detection to 0.795 average frame-level AUC across five unseen datasets.

desk verdict A plausible but under-specified deepfake detector whose headline cross-dataset gain rests on an unstated evaluation protocol. read the letter →

arxiv 2412.20799 v1 pith:7QWZ37TZ submitted 2024-12-30 cs.MM

Yuqi Li , Yuanzhong Zheng , Yaoxuan Wang , Jianjun Yin , Haojun Fei This is my paper · ORCID

classification cs.MM

keywords deepfakedetectionfeatureselectiondifferentialgeneexpressioncross-datasetgeneralizationfrequency-domainfeaturesdynamicweightingFaceForensics++LSTM

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

The paper asks whether a deepfake detector can do better by dynamically re-weighting its own feature channels depending on what kind of forgery it sees, rather than relying on static features. It proposes SFE-Net, a Selective Feature Expression Network that borrows the idea of differential gene expression: just as a cell activates different genes in different environments, the network activates different forensic cues for different inputs. On the standard protocol of training on FaceForensics++ (FF++ c23) and testing on five unseen deepfake datasets, SFE-Net reports 0.795 average frame-level AUC, above the best cited baseline SPSL at 0.767. The same module is shown to lift the fusion of five hand-crafted features from 0.771 to 0.795 in an ablation. A sympathetic reader would take the contribution as evidence that explicit, input-dependent feature selection is a viable path to cross-dataset generalization.

What carries the argument

The load-bearing object is the SFE-Net selection module itself, a small network placed after the five feature streams and before the classifier, which maps each input to a distribution of feature priorities, implemented with small convolutional branches, pooling, and softmax weighting. It plays the role of a gene-regulatory switch: the same set of forensic cues is available to the model at all times, but the module decides which cues are 'expressed' for the current frame. Its work is to convert five static feature maps into a context-dependent combination, and the ablation attributes the entire improvement from 0.771 to 0.795 to this module rather than to any one feature.

What would settle it

Re-running SFE-Net inside the same evaluation harness that produced the baseline numbers, with identical FF++ c23 training data, cropping, frame sampling, and AUC code, would settle the claim: if the average frame-level AUC over CDF-v1, CDF-v2, DFD, DFDC, and DFDCP falls to 0.767 or below, the reported cross-dataset margin disappears; if it stays near 0.795, the module's contribution is confirmed under the benchmark that generated the baseline results.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that inserting a selective feature expression module before a classifier lets a deepfake detector adapt its feature priorities to the synthesis method, and that this dynamic weighting transfers across datasets. The paper defines five forensic features, namely lighting consistency, high-frequency phase, compression reconstruction, morphological edge continuity, and texture, each extracted from preprocessed frames and fed through its own LSTM and softmax stream. SFE-Net then mixes these streams with a selection module, so the relevance of each cue is decided per input. Tables report that this design reaches 0.795 average frame-level AUC on Celeb-DF-v1, Celeb-DF-v2, DFD, DFDC, and DFDCP after training on FF++ c23, compared with 0.767 for the strongest cited baseline, and that removing SFE-Net drops the fusion of all five features to 0.771.

Load-bearing premise

The load-bearing premise is that SFE-Net was trained and evaluated under exactly the same protocol as the baselines, meaning the same FF++ c23 split, face preprocessing, frame sampling, and evaluation code, since the paper borrows baseline weight files from a benchmark but does not describe its own training details.

Editorial extensions

If this is right

Training on FF++ c23, SFE-Net reports frame-level AUC of 0.795 averaged over five unseen datasets, a margin of 0.028 over the best cited baseline SPSL.
Adding the selection module to the five-feature fusion raises average AUC from 0.771 to 0.795, so the paper attributes the gain to the module rather than to feature count.
The approach keeps an Xception backbone and standard FF++ training, meaning the reported improvement comes from changing the head of the pipeline, not from a new pretraining scheme.
On individual datasets the largest gains are on CDF-v1 at 0.866 and CDF-v2 at 0.798, while DFDC remains the weakest point at 0.709, which the authors state as a limitation.
Because the module is feature-agnostic, the same gating idea could be applied to other feature extractors without retraining the backbone from scratch.

Reading between the lines

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

If the reported protocol is exactly the one that produced the baseline numbers, namely the same FF++ c23 split, face preprocessing, frame sampling, and evaluation code, then a margin near 0.028 AUC is meaningful evidence that dynamic feature selection helps cross-dataset transfer; the paper never states that its own training and evaluation matched those baselines, so a reader should ask for that p
The biological analogy suggests a testable extension: inspect the learned feature weights on inputs from different synthesis families and check whether SFE-Net upweights phase and high-frequency cues on GAN-based forgeries and edge or morphological cues on warping-based ones.
The same gating idea could be ported to a single learned embedding instead of five hand-crafted features, yielding a test of whether the mechanism works by re-weighting alone rather than by the specific forensic features chosen here.
A stress test on heavily compressed or low-resolution videos would clarify whether the module's dynamic weighting remains useful when high-frequency and phase cues are largely destroyed.

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.

Desk Editor's Note

read the letter

Bottom line: this is a plausible incremental combination of known features for deepfake detection, but the central claim—that SFE-Net beats the best cited baseline by about 0.03 average AUC in cross-dataset testing—rests on an unspecified evaluation protocol. I can't tell whether the gain is real or a measurement artifact.

What's actually new: the architecture takes five handcrafted forensic features (light consistency, high frequency, compression artifacts, morphological ops, texture), runs each through its own LSTM stream, then uses a learned selection to combine them. That's a reasonable arrangement of existing components. The gene-expression analogy is just framing; it doesn't buy you a formal mechanism. The paper does report an ablation and is honest that the gain is modest and that DFDC remains hard.

The soft spots are serious. Section III.E says only that the baselines use DeepfakeBench weight files; it never says how SFE-Net was trained or evaluated. No optimizer, learning rate, batch size, frame count, face-crop details, or aggregation rule. Without that, Table I's 0.795 vs 0.767 is not a controlled comparison; face-crop padding and frame selection alone can move AUC by several points. The ablation in Table II has no error bars or repeated runs, and the selection mechanism is described only through a dense figure—no equations for the gating or the LSTM states. There's also no code or weights released. The 0.028 margin over SPSL is exactly the size of protocol sensitivity in this literature.

That said, the paper isn't incoherent. The method is describable, the datasets are standard, and the results are plausible. The flaw is under-reporting, not fabrication. If the authors can specify the protocol and release code, the claim becomes checkable.

Who it's for: anyone tracking incremental deepfake detection methods or using DeepfakeBench as a baseline. I wouldn't cite it now, but it's a reasonable warning example of why protocol reporting matters. I'd send it to review only with the explicit expectation that the missing details be supplied; in current form, it shouldn't be accepted.

Referee Report

4 major / 6 minor

Summary. The paper proposes SFE-Net, a deepfake detector that extracts five handcrafted features (lighting consistency, high-frequency/phase, compression-reconstruction, morphological, and texture cues) and uses a module of LSTM/softmax layers to adaptively weight these features before classification. Experiments trained on FF++ c23 and tested on Celeb-DF-v1, Celeb-DF-v2, DFD, DFDC, and DFDCP report an average frame-level AUC of 0.795, compared with 0.767 for the best baseline (SPSL), and an ablation showing that the SFE module improves a five-feature fusion from 0.771 to 0.795. The manuscript is largely descriptive: the biological analogy is elaborated at length, but the architecture, training protocol, and evaluation details are not specified beyond a figure and two short tables.

Significance. If the reported numbers were obtained under exactly the same protocol as the DeepfakeBench baselines, the result would be a modest but useful empirical contribution: training only on FF++ c23 and evaluating on five held-out datasets is the standard cross-dataset generalization setting, and a 0.028 average-AUC improvement over SPSL is practically interesting. The paper does not provide code, formal derivations, parameter-free constructions, or repeated-run statistics, so the contribution rests entirely on Tables I and II. The biological framing is not itself a technical contribution, and the proposed feature-selection mechanism is plausible but under-specified. The comparison against DeepfakeBench baselines is a good starting point and gives the paper a clear falsifiable empirical claim.

major comments (4)

[III.E, Table I] Section III.E / Table I: the headline comparison (SFE-Net 0.795 vs. SPSL 0.767 average frame-level AUC) is not a controlled comparison. The text states only that the baseline numbers use weight files provided by DeepfakeBench [26]; it never states the training split, face cropping/alignment, frame sampling, video-level aggregation, or evaluation code used for SFE-Net. In cross-dataset deepfake evaluation these choices can shift AUC by several points, so the reported margin may be an artifact of protocol rather than of the SFE-Net module. The authors should run SFE-Net inside the DeepfakeBench evaluation harness, or otherwise document the full protocol and demonstrate parity with the baselines.
[III.D, Table II] Section III.D / Table II: the ablation gain of 0.024 (0.771 to 0.795) is reported as a single number with no error bars, no number of runs, and no significance test. Since the variability of each row is not quantified, the claim that SFE-Net offers a consistent improvement over the static five-feature fusion is not supported. The authors should report mean and standard deviation over multiple random seeds and, ideally, per-dataset breakdowns for the ablation.
[III.C, Fig. 3] Section III.C / Fig. 3: the 'selective feature expression' mechanism is not formally defined. The text states only that each feature is fed into 'corresponding LSTM and softmax layers' and that the SFE-Net 'selectively enhances' features; no equations describe how the selection weights are computed, trained, or applied, nor how the LSTM outputs are combined with the landmark features in the classifier. Without this specification, the paper cannot be replicated and the claim that the module implements dynamic feature selection rather than a standard late-fusion classifier is unverifiable.
[IV (Conclusion)] Section IV: the conclusion acknowledges 'performance variance on complex datasets, relatively high computational costs, and preprocessing dependence' as limitations, but the experiments contain no measurements of variance, compute, or preprocessing sensitivity. These asserted limitations directly qualify the cross-dataset generalization claim and should be quantified or removed; as written, the text concedes that the robustness conclusion is untested.

minor comments (6)

[Fig. 3] Figure 3 uses two '(c)' subfigure labels ('(c) Feature Extraction' and '(c) SFE-Net'), and the subfigure labels in the caption do not match the surrounding text; this makes the architecture figure hard to follow.
[III.B] Section III.B says the paper reports video-level AUC, AP, and EER, but Table I contains only frame-level AUC; either add the promised metrics or delete the sentence.
[III.C, Table I] Table I lists SFE-Net with backbone 'Xception', but Section III.C does not mention Xception anywhere in the architecture description; the role of Xception (feature extractor, classifier, or both) needs to be stated.
[I, Fig. 1] The biological analogy is described at length, but no connection is made to any mathematical or algorithmic design choice; the analogy should be either operationalized or substantially reduced.
[Throughout] Several sentences are grammatically incomplete or mispunctuated (e.g., 'We show in Fig. 1. how to understand feature selective expression based on gene selective expression.' and 'In the SFE-Net, each of these features is fed into corresponding LSTM and softmax layers to further extract discontinuities between adjacent frames.'); a careful language edit is needed.
[III.A, Table I] The text refers to 'DeepFakeDetection [25]' while Table I uses the abbreviation 'DFDCP'; please make the dataset naming consistent.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper reports an empirical cross-dataset comparison with handcrafted features and learned fusion; the unstated SFE-Net evaluation protocol is a soundness issue, not a circular derivation.

full rationale

This paper makes an empirical claim rather than a derivational one. Its pipeline combines handcrafted feature extractors (morphological operations, lighting variance, Fourier phase, compression reconstruction, and LBP/GLCM texture features) with learned LSTM and softmax modules, trained on FF++ c23 and evaluated on held-out deepfake datasets. No equation in the paper is defined in terms of the target AUC, no parameter is fitted to the evaluation datasets and then renamed a prediction, and the baseline results are attributed to external weight files from DeepfakeBench [26]. The closest concern is that Section III.E describes how the baseline numbers were obtained but never specifies how SFE-Net itself was trained and evaluated, which makes the Table I comparison difficult to audit. That is a reproducibility and soundness issue, not circular reasoning. There is no self-citation chain, no imported uniqueness theorem, and no ansatz smuggled in via citation. Therefore no circular step can be exhibited, and the appropriate score is 0.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The paper contributes a network layout, but every one of its novelty claims depends on assumptions about feature informativeness, temporal modeling, transferability, and benchmark comparability. None of these assumptions is tested in isolation, and no code or data is shipped to check them. The free parameters are the learned SFE-Net weights and unreported training settings that drive the reported results.

free parameters (2)

SFE-Net selection and LSTM weights = not reported
The central claim rests on a learned feature-weighting module (Section III.C). No architecture dimensions, initialization, or trained weights are provided, so the mechanism that produces the reported AUC cannot be inspected or reproduced.
Training hyperparameters (learning rate, batch size, epochs, optimizer, input resolution, frame count) = not reported
These hand-chosen settings materially affect AUC and are absent from the paper, making the numbers in Table I impossible to reproduce.

assumptions (5)

domain assumption Morphological edge continuity, lighting variance, Fourier phase, compression artifacts, and texture statistics are informative for detecting the four deepfake families after FF++ training.
Section II assigns one feature to each manipulation type but gives no quantitative evidence that these features are discriminative on the test datasets.
domain assumption Feeding each feature stream through its own LSTM and softmax implements meaningful temporal and cross-feature adaptation rather than a fixed fusion.
Section III.C describes the pipeline at a high level; no evidence or ablation separates temporal modeling from the claimed adaptive selection.
domain assumption Training only on FF++ c23 transfers to Celeb-DF, DFD, DFDC, and DFDCP under the stated AUC.
Cross-dataset generalization is the core claim, but no train-test distribution analysis, domain adaptation, or protocol details are given in Section III.E.
domain assumption Baseline results from DeepfakeBench [26] were produced under a protocol equivalent to SFE-Net's own evaluation.
Section III.E mentions baseline weight files from [26] but does not document SFE-Net's preprocessing, frame sampling, thresholding, or aggregation.
standard math Standard definitions of erosion, dilation, opening, Fourier transform, LBP, and GLCM are valid image-processing tools.
Sections II.A-II.D invoke these classical operations without proof; this is acceptable background.

how reviews work

0 comments

Cite this review

Pith. "Pith review of SFE-Net: Harnessing Biological Principles of Differential Gene Expression for Improved Feature Selection in Deep Learning Networks." pith.science (2026). https://pith.science/paper/7QWZ37TZ

@misc{pith2026241220799,
  author       = {Pith},
  title        = {Pith review of: SFE-Net: Harnessing Biological Principles of Differential Gene Expression for Improved Feature Selection in Deep Learning Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7QWZ37TZ}},
  note         = {Machine review of arXiv:2412.20799}
}

read the original abstract

In the realm of DeepFake detection, the challenge of adapting to various synthesis methodologies such as Faceswap, Deepfakes, Face2Face, and NeuralTextures significantly impacts the performance of traditional machine learning models. These models often suffer from static feature representation, which struggles to perform consistently across diversely generated deepfake datasets. Inspired by the biological concept of differential gene expression, where gene activation is dynamically regulated in response to environmental stimuli, we introduce the Selective Feature Expression Network (SFE-Net). This innovative framework integrates selective feature activation principles into deep learning architectures, allowing the model to dynamically adjust feature priorities in response to varying deepfake generation techniques. SFE-Net employs a novel mechanism that selectively enhances critical features essential for accurately detecting forgeries, while reducing the impact of irrelevant or misleading cues akin to adaptive evolutionary processes in nature. Through rigorous testing on a range of deepfake datasets, SFE-Net not only surpasses existing static models in detecting sophisticated forgeries but also shows enhanced generalization capabilities in cross-dataset scenarios. Our approach significantly mitigates overfitting by maintaining a dynamic balance between feature exploration and exploitation, thus producing more robust and effective deepfake detection models. This bio-inspired strategy paves the way for developing adaptive deep learning systems that are finely tuned to address the nuanced challenges posed by the varied nature of digital forgeries in modern digital forensics.

Figures

Figures reproduced from arXiv: 2412.20799 by the authors.

**Figure 2.** Visualizing the comparison of different feature extractions. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** SFE-Net architecture where F denotes the Fourier Transform of the spatial domain image f(x, y). The high-pass filter H(u, v) is defined by: H(u, v) = ( 0 if (u, v) is central 1 otherwise Phase Extraction and Image Reconstruction: The phase spectrum ϕ(u, v) is extracted for reconstruction: ϕ(u, v) = arg(F ′ (u, v)), G(u, v) = e iϕ(u,v) , g(x, y) = F −1 {G(u, v)} This compact approach emphasizes phase information to e… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 22 canonical work pages

[26]

Z. Yan, Y . Zhang, X. Yuan, S. Lyu, and B. Wu, ”Deep- fakeBench: A comprehensive benchmark of deepfake detection,” in Advances in Neural Information Processing Systems, vol. 36, pp. 4534–4565, 2023. © 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including r...

work page 2023
[1]

Branwen and S

G. Branwen and S. Presser, ”GPT-2 neural network poetry,” Gwern.net, vol. 29, Oct. 2019

work page 2019
[2]

Westerlund, ”The emergence of deepfake technology: A review,” Technology Innovation Management Review , vol

M. Westerlund, ”The emergence of deepfake technology: A review,” Technology Innovation Management Review , vol. 9, no. 11, 2019

work page 2019
[3]

Vincent, ”Watch Jordan Peele use AI to make Barack Obama deliver a PSA about fake news,” The V erge, vol

J. Vincent, ”Watch Jordan Peele use AI to make Barack Obama deliver a PSA about fake news,” The V erge, vol. 17, pp. 1, 2018

work page 2018
[4]

Suwajanakorn, S

S. Suwajanakorn, S. M. Seitz, and I. Kemelmacher- Shlizerman, ”Synthesizing Obama: Learning lip sync from audio,” ACM Transactions on Graphics (ToG) , vol. 36, no. 4, pp. 1–13, 2017, ACM New York, NY , USA

work page 2017
[5]

Fallis and K

D. Fallis and K. Mathiesen, ”Fake news is counterfeit news,” Inquiry, pp. 1–20, 2019, Taylor & Francis

work page 2019
[6]

De Ruiter, ”The distinct wrong of deepfakes,” Philos- ophy & Technology, vol

A. De Ruiter, ”The distinct wrong of deepfakes,” Philos- ophy & Technology, vol. 34, no. 4, pp. 1311–1332, 2021, Springer

work page 2021
[7]

Vaccari and A

C. Vaccari and A. Chadwick, ”Deepfakes and disinforma- tion: Exploring the impact of synthetic political video on deception, uncertainty, and trust in news,” Social Media+ Society, vol. 6, no. 1, pp. 2056305120903408, 2020, SAGE Publications

work page 2020

Show all 26 references

[8]

Afchar, V

D. Afchar, V . Nozick, J. Yamagishi, and I. Echizen, ”Mesonet: A compact facial video forgery detection network,” in 2018 IEEE International Workshop on In- formation F orensics and Security (WIFS), pp. 1–7, 2018, IEEE

2018
[9]

K. He, X. Zhang, S. Ren, and J. Sun, ”Deep resid- ual learning for image recognition,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770–778, 2016

2016
[10]

Rossler, D

A. Rossler, D. Cozzolino, L. Verdoliva, C. Riess, J. Thies, and M. Nießner, ”Faceforensics++: Learning to detect manipulated facial images,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1–11, 2019

2019
[11]

Tan, ”Efficientnet: Rethinking model scaling for convolutional neural networks,” arXiv preprint arXiv:1905.11946, 2019

M. Tan, ”Efficientnet: Rethinking model scaling for convolutional neural networks,” arXiv preprint arXiv:1905.11946, 2019

1905 arXiv
[12]

H. H. Nguyen, J. Yamagishi, and I. Echizen, ”Capsule- forensics: Using capsule networks to detect forged im- ages and videos,” in ICASSP 2019-2019 IEEE Inter- national Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 2307–2311, 2019, IEEE

2019
[13]

Li and S

Y . Li and S. Lyu, ”Exposing deepfake videos by detecting face warping artifacts,” arXiv preprint arXiv:1811.00656, 2018

2018 arXiv
[14]

L. Li, J. Bao, T. Zhang, H. Yang, D. Chen, F. Wen, and B. Guo, ”Face x-ray for more general face forgery detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 5001– 5010, 2020

2020
[15]

H. Dang, F. Liu, J. Stehouwer, X. Liu, and A. K. Jain, ”On the detection of digital face manipulation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 5781–5790, 2020

2020
[16]

Y . Ni, D. Meng, C. Yu, C. Quan, D. Ren, and Y . Zhao, ”Core: Consistent representation learning for face forgery detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 12–21, 2022

2022
[17]

J. Cao, C. Ma, T. Yao, S. Chen, S. Ding, and X. Yang, ”End-to-end reconstruction-classification learn- ing for face forgery detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4113–4122, 2022

2022
[18]

Z. Yan, Y . Zhang, Y . Fan, and B. Wu, ”UCF: Uncovering common features for generalizable deepfake detection,” in Proceedings of the IEEE/CVF International Confer- ence on Computer Vision , pp. 22412–22423, 2023

2023
[19]

Y . Qian, G. Yin, L. Sheng, Z. Chen, and J. Shao, ”Thinking in frequency: Face forgery detection by min- ing frequency-aware clues,” in European Conference on Computer Vision, pp. 86–103, 2020, Springer

2020
[20]

H. Liu, X. Li, W. Zhou, Y . Chen, Y . He, H. Xue, W. Zhang, and N. Yu, ”Spatial-phase shallow learning: Re- thinking face forgery detection in frequency domain,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 772–781, 2021

2021
[21]

Y . Luo, Y . Zhang, J. Yan, and W. Liu, ”Generalizing face forgery detection with high-frequency features,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 16317–16326, 2021

2021
[22]

Y . Li, X. Yang, P. Sun, H. Qi, and S. Lyu, ”Celeb-DF: A large-scale challenging dataset for deepfake forensics,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 3207–3216, 2020

2020
[23]

Dolhansky, R

B. Dolhansky, R. Howes, B. Pflaum, N. Baram, and C. C. Ferrer, ”The deepfake detection challenge (DFDC) preview dataset,” arXiv preprint arXiv:1910.08854, 2019

1910 arXiv
[24]

Dolhansky, J

B. Dolhansky, J. Bitton, B. Pflaum, J. Lu, R. Howes, M. Wang, and C. C. Ferrer, ”The deepfake de- tection challenge (DFDC) dataset,” arXiv preprint arXiv:2006.07397, 2020

2006 arXiv
[25]

Google AI Blog, ”Contributing data to deepfakedetec- tion,” Accessed: 2021-11-13

2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [26]

Z. Yan, Y . Zhang, X. Yuan, S. Lyu, and B. Wu, ”Deep- fakeBench: A comprehensive benchmark of deepfake detection,” in Advances in Neural Information Processing Systems, vol. 36, pp. 4534–4565, 2023. © 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including r...

work page 2023

[2] [1]

Branwen and S

G. Branwen and S. Presser, ”GPT-2 neural network poetry,” Gwern.net, vol. 29, Oct. 2019

work page 2019

[3] [2]

Westerlund, ”The emergence of deepfake technology: A review,” Technology Innovation Management Review , vol

M. Westerlund, ”The emergence of deepfake technology: A review,” Technology Innovation Management Review , vol. 9, no. 11, 2019

work page 2019

[4] [3]

Vincent, ”Watch Jordan Peele use AI to make Barack Obama deliver a PSA about fake news,” The V erge, vol

J. Vincent, ”Watch Jordan Peele use AI to make Barack Obama deliver a PSA about fake news,” The V erge, vol. 17, pp. 1, 2018

work page 2018

[5] [4]

Suwajanakorn, S

S. Suwajanakorn, S. M. Seitz, and I. Kemelmacher- Shlizerman, ”Synthesizing Obama: Learning lip sync from audio,” ACM Transactions on Graphics (ToG) , vol. 36, no. 4, pp. 1–13, 2017, ACM New York, NY , USA

work page 2017

[6] [5]

Fallis and K

D. Fallis and K. Mathiesen, ”Fake news is counterfeit news,” Inquiry, pp. 1–20, 2019, Taylor & Francis

work page 2019

[7] [6]

De Ruiter, ”The distinct wrong of deepfakes,” Philos- ophy & Technology, vol

A. De Ruiter, ”The distinct wrong of deepfakes,” Philos- ophy & Technology, vol. 34, no. 4, pp. 1311–1332, 2021, Springer

work page 2021

[8] [7]

Vaccari and A

C. Vaccari and A. Chadwick, ”Deepfakes and disinforma- tion: Exploring the impact of synthetic political video on deception, uncertainty, and trust in news,” Social Media+ Society, vol. 6, no. 1, pp. 2056305120903408, 2020, SAGE Publications

work page 2020