-->PONE-D-25-56123-->-->STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization-->-->PLOS One

Dear Dr. Mu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.-->--> -->-->Please revise your manuscript in accordance with the reviewers’ comments-->-->

Please submit your revised manuscript by Feb 12 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

• A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Fatih Uysal, Ph.D.

Academic Editor

PLOS One

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please note that PLOS One has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, we expect all author-generated code to be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

3. Thank you for stating in your Funding Statement:

[This work was supported in part by the State Grid Information and Telecommunication

Group Co., Ltd., which coordinates scientific and technological projects (SGIT0000XTJS2401078).].

Please provide an amended statement that declares *all* the funding or sources of support (whether external or internal to your organization) received during this study, as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now. Please also include the statement “There was no additional external funding received for this study.” in your updated Funding Statement.

Please include your amended Funding Statement within your cover letter. We will change the online submission form on your behalf.

4. Thank you for stating the following in your manuscript:

[This work was supported in part by the State Grid Information and Telecommunication

Group Co., Ltd., which coordinates scientific and technological projects (SGIT0000XTJS2401078).]

We note that you have provided funding information that is currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

[This work was supported in part by the State Grid Information and Telecommunication

Group Co., Ltd., which coordinates scientific and technological projects (SGIT0000XTJS2401078).]

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

5. Thank you for stating the following in the Financial Disclosure section:

[This work was supported in part by the State Grid Information and Telecommunication

Group Co., Ltd., which coordinates scientific and technological projects (SGIT0000XTJS2401078).].

We note that you received funding from a commercial source: State Grid Information and Telecommunication Group Co., Ltd.

Please provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other relevant declarations relating to employment, consultancy, patents, products in development, marketed products, etc.

Within this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf.

6. Thank you for stating the following in the Competing Interests section:

[The authors have declared that no competing interests exist.].

We note that one or more of the authors are employed by a commercial company State Grid Information and Telecommunication Group Co., Ltd.

1. Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form.

Please also include the following statement within your amended Funding Statement.

“The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

2. Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

Within your Competing Interests Statement, please confirm that this commercial affiliation does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests) . If this adherence statement is not accurate and there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

7. In the online submission form, you indicated that [The code implementation of STGAD used for the experiments is available from the corresponding author upon reasonable request.].

All PLOS journals now require all data underlying the findings described in their manuscript to be freely available to other researchers, either 1. In a public repository, 2. Within the manuscript itself, or 3. Uploaded as supplementary information.

This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If your data cannot be made publicly available for ethical or legal reasons (e.g., public availability would compromise patient privacy), please explain your reasons on resubmission and your exemption request will be escalated for approval.

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

Additional Editor Comments:

Please revise your manuscript in accordance with the reviewers’ comments.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: Yes

Reviewer #2: Partly

**********

-->2. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

Reviewer #2: No

**********

-->5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: The manuscript titled “STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization” presents a generative adversarial approach enhanced with a Self-Temporal Transformer component and dual scoring mechanism. The topic is highly relevant to multivariate anomaly detection in cyber-physical systems, and the authors conduct extensive evaluations on several widely used datasets. The work shows promising empirical performance and includes ablation studies, robustness tests, and interpretability analysis.

However, while the contribution is meaningful, the manuscript requires major revision to improve clarity, methodological rigor, and overall presentation.

Major Strengths

The introduction and related work provide a clear overview of the challenges in multivariate time-series anomaly detection.

Integrating global self-attention via a Self-Temporal Transformer is technically interesting and appears effective.

Evaluations are conducted across five benchmark datasets (SMD, SMAP, MSL, SWaT, MIT-BIH), along with ablations and noise robustness tests.

The inclusion of attention visualization and statistical deviation alignment contributes positively to transparency.

Major Concerns Requiring Revision

1.The manuscript is significantly longer than typical research articles, with repeated explanations of datasets, windowing strategy, and model design.The writing includes grammatical errors, inconsistent notation, and overly complex phrasing. A substantial readability-focused revision is necessary.

2. The generator is non-conditional, yet reconstruction errors are computed by sampling multiple noise vectors and taking a minimum.This raises several questions:

Why is a non-conditional generator preferred over a conditional reconstruction model?

How is the minimum reconstruction error conceptually justified?

How many samples are generated per window, and does this introduce variance at inference?

This component needs clearer explanation and justification.

3. A window length of 5 timesteps is unusually small for datasets like SMAP, MSL, SWaT, and SMD, where long temporal dependencies are known to be significant.Using such short windows may weaken baselines designed for longer sequences.A justification and/or sensitivity study on window size is needed.

4. Although the authors state that baselines were trained under identical conditions, more detail is required:

Were all models given identical preprocessing steps, window sizes, and normalization?

Was the same thresholding method (e.g., POT) applied to all baselines?

Were baseline hyperparameters tuned fairly?

These details are essential to ensure a fair comparison.

5. Given the computational demands of Transformers and GANs, it is important to include:Training time per epoch,Inference throughput,GPU memory usage,Complexity scaling with number of variables.This is particularly relevant for real-world industrial deployment.

6.Interpretability section is overly detailed.While valuable, the section is disproportionately long and contains case-specific minutiae that obscure key insights. A more concise, structured presentation would improve readability.

Minor Issues

Some equations (e.g., Eq. (4)) contain additional terms not clearly motivated.

Hyperparameter β is inconsistently described in different sections (0.1 vs 0.7).

Several figures are low-resolution or too small to interpret easily.

In ablation tables, entries such as "<0.01" are unclear—are these true values or placeholders?

Formatting inconsistencies (e.g., leftover LaTeX commands, alignment issues) should be corrected.

Reviewer #2: The manuscript proposes STGAD, a novel unsupervised anomaly detection framework for multivariate time series that integrates a Generator–Discriminator architecture (based on WGAN-GP) with a Self-Temporal Transformer in the discriminator and a dual-score fusion strategy. The paper is well-motivated, addresses a timely problem in AI for cyber-physical systems, and reports strong empirical results across five benchmark datasets. The use of global self-attention, WGAN-GP for stability, and POT-based adaptive thresholding are reasonable design choices. However, several significant issues, ranging from methodological ambiguity and inconsistent experimental details to overstatement of claims, need to be addressed before the work can be considered for publication.

1. In Section 3.3, the fusion weight β is described as a tunable hyperparameter. But in Section 4.2, the authors fix

β=0.7 (or earlier they say “0.1 reconstruction / 0.9 discriminator”). No justification is given for this choice, and it is unclear whether this was tuned per dataset or globally fixed. If fixed globally, why does the discriminator dominate? If tuned per dataset, this undermines the “unsupervised” and “transferable” claims—since tuning requires validation labels or proxy metrics.

2. The paper claims the Generator is “non-conditional,” yet during inference it appears to reconstruct real inputs. This contradiction needs clarification: Is STGAD a pure GAN, or a hybrid reconstruction+adversarial model? If the latter (which seems to be the case), the architecture is more akin to USAD or MAD-GAN than a standard GAN, and this should be explicitly acknowledged.

3. The Self-Temporal Transformer is described as applying “global within-window self-attention,” but it’s unclear how this differs from standard Transformer encoders (e.g., in TranAD). What is "self-temporal" beyond standard temporal attention? The term appears to be coined without sufficient technical distinction.

4. The ablation study shows catastrophic failure of the "Vanilla GAN" variant (F1 ≈ 0 on most datasets), which, while expected, suggests that WGAN-GP is doing most of the work, not the Transformer. The paper attributes success to the “Self-Temporal Transformer,” but the ablation shows that removing attention causes only minor drops (e.g., SMD F1: 0.9756 → 0.8582), whereas removing WGAN-GP destroys performance entirely. This undermines the claimed novelty.

5. The paper uses “MBA” to refer to the MIT-BIH Arrhythmia database, but the standard acronym is MITDB or MIT-BIH. “MBA” typically refers to “Master of Business Administration,” which is confusing.

6. Numerous formatting issues (e.g., line breaks in mid-sentence, inconsistent spacing: “main-taining”, “Unsupe rvised”, “p rep rint”) suggest the manuscript was not carefully proofread.

7. While the case study on SMAP is visually compelling, the Saliency–Local and Saliency–Global overlaps are all reported as 1.0, which is statistically implausible and suggests post-hoc cherry-picking or overly generous Top-K matching (e.g., K may be too large relative to D).

8. The paper claims WGAN-GP improves stability, but no metrics (e.g., discriminator loss variance, mode collapse diagnostics) are provided to support this beyond final F1 scores.

9. STGAD is tested under higher noise levels (σ = 0.5) while baselines are only tested up to σ = 0.25. This makes comparisons biased in favor of STGAD. A fair comparison would test all methods at identical noise intensities.

**********

-->6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: Yes: Amanul Islam

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

-->PONE-D-25-56123R1-->-->STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization-->-->PLOS One

Dear Dr. Mu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.-->--> -->-->Reviewer 2 has some methodological concerns that warrant a response. -->-->

Please submit your revised manuscript by May 07 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

• A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Daniel Parkes, PhD

Staff Editor

PLOS One

Journal Requirements:

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.-->

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

-->2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

-->3. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

-->4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

Reviewer #2: No

**********

-->6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: The manuscript titled “STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization” proposes a hybrid GAN-based framework that integrates a self-temporal transformer within a WGAN-GP discriminator for anomaly detection in multivariate time-series. The model combines reconstruction residuals with discriminator confidence and applies POT-based thresholding for anomaly detection. Experimental results on five benchmark datasets (SMD, SMAP, MSL, SWaT, MIT-BIH) show improved performance compared to several baseline methods.

Overall, the paper addresses an important problem in anomaly detection and presents a technically sound framework with promising results.

Strengths

1.The paper addresses a relevant and important problem in multivariate time-series anomaly detection.

2.The proposed framework effectively combines GAN-based modeling and Transformer-based temporal attention.

3.The study includes extensive experiments on multiple benchmark datasets, demonstrating competitive performance.

4.The authors provide ablation and stability analyses to justify the effectiveness of the model components.

4.The manuscript is generally well structured and clearly organized.

Weaknesses / Points for Improvement

1.Model novelty should be clarified further. The Self-Temporal Transformer appears to be a standard Transformer encoder integrated into the discriminator, so the novelty compared with existing transformer-based anomaly detection models should be more clearly emphasized.

2.Generator design justification is limited. The use of a non-conditional generator with sample matching during inference may introduce additional computational overhead. The authors should discuss the efficiency and scalability of this approach in practical deployments.

3.Window size selection (L=5) appears relatively small. More justification or broader sensitivity analysis would strengthen the methodological choices.

4.Interpretability remains limited. Although the authors include a short interpretability discussion, further insights into how anomalies are localized across variables would improve the practical usefulness of the method.

5.Minor language and formatting issues remain and should be carefully proofread before publication.

The manuscript presents a technically sound and well-evaluated approach for multivariate time-series anomaly detection. With minor revisions addressing the points above, the paper could be suitable for publication.

Reviewer #2: The author's response has addressed my previous concerns. I recommend further polishing the English writing before formal publication. Additionally, the resolution of the figures is currently low, which affects the reading experience; I suggest replacing all figures with vector formats.

**********

-->7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: Yes: Amanul Islam

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

-->PONE-D-25-56123R2-->-->STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization-->-->PLOS One

Dear Dr. Mu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by May 30 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

• A letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

-->

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

As the corresponding author, your ORCID iD is verified in the submission system and will appear in the published article. PLOS supports the use of ORCID, and we encourage all coauthors to register for an ORCID iD and use it as well. Please encourage your coauthors to verify their ORCID iD within the submission system before final acceptance, as unverified ORCID iDs will not appear in the published article. Only the individual author can complete the verification step; PLOS staff cannot verify ORCID iDs on behalf of authors.

We look forward to receiving your revised manuscript.

Kind regards,

Abdul Ahad

Academic Editor

PLOS One

Journal Requirements:

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.-->

Reviewer #1: All comments have been addressed

Reviewer #3: All comments have been addressed

**********

-->2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->3. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

Reviewer #3: No

**********

-->5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: The revised manuscript presents a well-structured and technically sound framework for unsupervised multivariate time-series anomaly detection. The authors have made substantial improvements in response to prior comments, particularly in clarifying the methodological contributions, strengthening experimental analysis, and improving overall presentation quality.

The clarification regarding the role of the Transformer component is appropriate, and it is now clear that the novelty lies in the overall framework design rather than in proposing a new attention mechanism. The addition of efficiency analysis is valuable and provides useful insights into the scalability and practical deployment of the proposed approach. The sensitivity analysis on window length and other hyperparameters further strengthens the methodological justification and enhances reproducibility.

The expanded interpretability analysis is a notable improvement. Including both qualitative and quantitative evaluations (e.g., overlap-based metrics) provides better evidence that the model can localize anomalies across variables in a meaningful way. This significantly improves the practical relevance of the work.

The experimental evaluation is comprehensive, covering multiple benchmark datasets and including ablation, robustness, and sensitivity studies. The results consistently demonstrate the effectiveness of the proposed dual-score mechanism and the stability benefits of WGAN-GP. The inclusion of computational cost analysis (training time and inference latency) is also appreciated, as it highlights the feasibility of the method in near-real-time scenarios.

Despite these improvements, a few minor issues remain:

1.While interpretability has been improved, the explanation of how practitioners can directly use the localization results in real-world applications could be further elaborated.

2.The discussion on the trade-off between sampling size (N) and inference cost could be made more explicit, especially for strict real-time deployment scenarios.

3.Some minor language and grammatical issues may still remain and could benefit from one final round of careful proofreading.

Overall, the manuscript is significantly improved and addresses the major concerns raised in previous rounds. The proposed method is well-motivated, experimentally validated, and practically relevant.

Reviewer #3: The research paper titled “STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization” discusses a research area of significant interest to the scientific community. However, here are few suggestions to improve the quality of the manuscript:

For suggestion 1 to 2, author can refer and cite the paper titled “Multilingual entity alignment by abductive knowledge reasoning on multiple knowledge graphs” (https://doi.org/10.1016/j.engappai.2024.109660) at a suitable location.

1. It is suggested to briefly discuss the structure of subsequent sections in a single paragraph at the end of introduction section.

2. It is suggested to include the DOI information for each research paper in the reference section.

**********

-->7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: Yes: Amanul Islam

Reviewer #3: Yes: Muhammad Usman Akhtar

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation.

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

-->

STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization

PONE-D-25-56123R3

Dear Dr. Mu,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Academic Editor

PLOS One

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.-->

Reviewer #1: All comments have been addressed

Reviewer #3: All comments have been addressed

**********

-->2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->3. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

Reviewer #3: Yes

**********

-->6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: The authors present a revised version of the manuscript entitled “STGAD: Self-temporal generative adversarial framework with transformer attention for unsupervised multivariate time-series anomaly detection and localization.” The paper addresses an important problem in multivariate time-series anomaly detection and proposes a hybrid framework combining GAN-based modeling with Transformer-based temporal representation and dual-score fusion.

Overall, the revised manuscript shows clear improvement compared to the previous version. The authors have adequately addressed most of the reviewer comments, particularly by enhancing the explanation of practical applicability, clarifying the trade-off between sampling size and inference cost, and improving the overall organization and readability of the paper. The added discussion on interpretability and deployment scenarios is helpful and strengthens the practical relevance of the work.

The technical contribution is solid, and the experimental evaluation is comprehensive, covering multiple benchmark datasets and including ablation, robustness, and sensitivity analyses. The dual-score mechanism and the integration of Transformer-based temporal modeling within a WGAN-GP framework are well-motivated and demonstrate consistent performance improvements over baseline methods.

However, a few minor issues still remain:

1. While the authors improved the discussion, the real-world deployment pipeline could still be described more concretely (e.g., integration with monitoring systems, computational constraints in edge environments).

2.The manuscript would benefit from a slightly more explicit discussion comparing the added architectural complexity with the achieved performance gains, especially for practitioners considering implementation.

3.Although datasets and code are mentioned, it would be helpful to ensure that all implementation details (e.g., exact hyperparameter settings for each dataset, random seeds, and preprocessing nuances) are fully specified for easier reproducibility.

4.The manuscript is generally well-written, but a final round of proofreading could further improve fluency and remove a few remaining minor grammatical inconsistencies.

From an ethical and publication standpoint, I did not identify concerns related to plagiarism, dual publication, or research ethics. The data sources and funding disclosures appear appropriate and transparent.

Finally, the manuscript is technically sound, relevant, and significantly improved after revision. With minor refinements as suggested above, it is suitable for publication.

Reviewer #3: The author has addressed all of my suggestions; therefore, we believe that the quality of the manuscript has improved and recommend it for publication.

**********

-->7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.-->

Reviewer #1: Yes: Amanul Islam

Reviewer #3: Yes: Muhammad Usman Akhtar

**********