PONE-D-24-50383An Attribute-Enhanced Relationship-Aware Neighborhood Matching Model with Dual AttentionPLOS ONE

Dear Dr. YANG,

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 Mar 22 2025 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 rebuttal 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,

Tao Huang

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, all author-generated code must 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 the following in the Acknowledgments Section of your manuscript:

“This work was supported by the Jiangsu Provincial Department of Industry and Information Technology Key Technology Innovation Project Guidance Plan under Grant 141-62-65, the Digital Public Service Platform Project of Jiangsu Provincial Department of Science and Technology under Grant 93208000931, the Industry-university-research Project of Jiangsu Provincial Department of Science and Technology under Grant 20221343, and Fujian Provincial Financial Research General Funding Project under Grant 2023CZ50.”

We note that you have provided funding information that is not 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:

“Junlin Gu is fundered by the Jiangsu Provincial Department of Industry and Information Technology Key Technology Innovation Project Guidance Plan under Grant 141-62-65, the Digital Public Service Platform Project of Jiangsu Provincial Department of Science and Technology under Grant 93208000931, and the Industry-university-research Project of Jiangsu Provincial Department of Science and Technology under Grant 20221343. Xiong Yang is fundered by Fujian Provincial Financial Research General Funding Project under Grant 2023CZ50. All funders play an important role in the study design, data collection and analysis, decision to publish, and preparation of the manuscript.”

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

[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: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #5: Yes

Reviewer #6: Partly

Reviewer #7: Yes

**********

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

Reviewer #4: Yes

Reviewer #5: Yes

Reviewer #6: Yes

Reviewer #7: 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

Reviewer #3: No

Reviewer #4: Yes

Reviewer #5: Yes

Reviewer #6: Yes

Reviewer #7: 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: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #5: Yes

Reviewer #6: Yes

Reviewer #7: Yes

**********

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 authors propose ARNM-DAE2A, a model that combines four modules: an entity structure embedding module, an attribute structure embedding module, a joint alignment module, and a relationship-aware neighborhood matching module. The model aims to enhance the structural learning capability of GCN by introducing a pairwise attention mechanism, utilize GCN for acquiring entity attribute information, and fuse relationship and attribute information to create a comprehensive representation of entities. The relationship-aware neighborhood matching module corrects noise in the aggregated information by comparing the neighborhood relationships of entity pairs. The authors evaluate their model on three DBP15K datasets and demonstrate superior performance compared to several baseline models. There are some small issues should be addressed:

1. Suggest adding some references published in the past two years.

2. Suggest adding some advanced baselines for comparison.

3. The format of the table in the manuscript seems to be incorrect, it is recommended to check.

4. There are still some typos in the manuscript.

Reviewer #2: Summary:

The manuscript proposes ARNM-DAE2A, a novel entity alignment model that enhances knowledge graph (KG) integration through dual attention mechanisms and attribute information. It combines four modules: entity structure embedding, attribute structure embedding, joint alignment, and relationship-aware neighborhood matching. Experiments on the DBP15K datasets demonstrate state-of-the-art performance, significantly outperforming baselines.

Strengths:

- Novelty: The proposed ARNM-DAE2A model effectively integrates attribute information and a dual attention mechanism, which improves structural learning in graph convolutional networks (GCNs) for entity alignment tasks.

- Performance: Experimental results demonstrate significant improvements over existing baseline models on standard datasets (DBP15K), with detailed metrics (e.g., Hits@K, MRR) supporting these claims.

- Thoroughness: The manuscript includes well-explained methodologies, ablation studies, and hyperparameter analyses that validate the model's design.

- Clarity: The writing is clear and logical, with figures and tables aiding comprehension.

Weaknesses:

- Innovation Depth: While the dual attention mechanism and attribute integration are valuable, the innovations appear incremental compared to existing GCN-based approaches.

- Evaluation Diversity: The datasets (DBP15K) used for evaluation, while standard, are limited in scope. Broader evaluations across different types of knowledge graphs (e.g., domain-specific KGs) would strengthen the claims.

- Reproducibility: Although the experimental setup is detailed, additional information on hyperparameters and code availability would enhance reproducibility.

- Comparison Baselines: While key baselines are included, some more recent models in entity alignment using advanced techniques like transformers or hyperbolic embeddings could have been added for a stronger comparative evaluation.

Recommendation:

I recommend the manuscript for major revision with the following suggestions:

- Expand Evaluation: Include experiments on more diverse and domain-specific datasets.

- Add Baselines: Compare the model against newer methods, such as transformer-based or hyperbolic embedding approaches.

- Reproducibility: Provide a link to the implementation or specify more details about the experimental environment.

- Highlight Practical Use Cases: Discuss potential real-world applications or case studies where this model could be impactful.

I hope my review can improve your work!

Reviewer #3: The paper proposes an entity alignment model named, ARNM-DAE2A. This model integrates both structural and attribute information and introduces a dual attention mechanism to optimize the parameters. Experimental results demonstrate that the proposed method achieves SOTA performance.

Strengths

- The paper is overall well-written.

- The motivation for introducing the dual attention mechanism is well-discussed.

- The authors thoroughly review existing entity alignment methods.

- Experimental results show that the proposed model outperforms baselines.

Weaknesses

- The abstract could be improved by explicitly discussing the motivation for using dual attention mechanisms and the significance of integrating attribute information.

- Despite emphasizing the role of attributes in the introduction, the method simply uses attributes as GNN inputs rather than actively leveraging them in interaction with the structural embeddings.

- Some hyperparameters (e.g., seed alignment ratio and hyperparameters in Equation 17) are arbitrarily set without detailed discussion or justification.

- The experiments are conducted only on the DBP15K dataset. The model could be potentially evaluated on other, more diverse datasets (e.g., Wikidata, YAGO) as well.

- In the ablation experiments, some configurations (e.g., MM+AE) outperform the final model and may require some discussions or explanations.

- The model is quite complex. Discussions or empirical analysis on its scalability are required.

- The baselines used in the experiments are relatively old. The most recent baseline is from 2020. There are more recent methods that can be compared as baselines.

- A minor typo in Table 1 "tripes."

- The source code is not shared.

In summary, the paper makes a meaningful contribution to the field of entity alignment by proposing a dual attention mechanism and relationship-aware neighborhood matching model. However, there are some limitations on the method and experiments as discussed above. Thus, I recommend a revision of this paper.

Reviewer #4: This paper is a valuable contribution to the field of knowledge graph entity alignment. The methodology is robust, and the results are promising. However, with some additional effort, the manuscript could reach its full potential and make an even greater impact. I recommend a Minor Revision to address these areas of improvement.

While the manuscript is well-written in standard English, some sections could benefit from clearer and more concise explanations. For instance, the description of the dual attention mechanism, while technically sound, might be challenging for readers unfamiliar with the concept. Including illustrative examples or visual aids could greatly enhance comprehension and accessibility.

Additionally, the practical implications and potential real-world applications of the proposed method deserve more attention. Expanding on how this approach could be applied to larger or more heterogeneous knowledge graphs, or how it handles noisy or incomplete data, would highlight its broader relevance and strengthen its appeal to a wider audience.

These refinements, though requiring additional effort, will significantly enhance the clarity, accessibility, and impact of this important work. I appreciate the dedication already invested in this manuscript and encourage the authors to take this opportunity to further elevate its quality and reach.

Reviewer #5: The authors propose a fusion of several established techniques for assessing entity alignment within Knowledge Graphs (KGs). They iteratively apply GCNs and RDGCNs to KGs and associated attribute networks in a principled manner in order to optimize relationship mapping. The result is a notable increase in alignment performance using established public datasets and metrics.

The methodology appears sound and is rigorously explained.

There are several minor recommendations or questions I can offer, primarily to improve understandability.

First and foremost, I recommend the authors publish their code in a git repository such as GitHub. This would substantially improve usability of their model for other researchers.

The mathematical formulation of the model is rigorously documented, but the relation between the notation and Figure 1 seems incompletely explored. A short explanation which relates the figure back to the notation could enhance readability for readers unfamiliar with the problem domain. Similarly, it may help to clarify what the characters on the left-hand-side of Figure 1 mean for readers who do not know the language. As someone who does not speak or read what I presume to be Mandarin, I can only assume that the entities on the left hand side are direct translations of those same entities on the right hand side.

In section 5.1, I believe that it could be helpful to note that the train/test split used was an exact match to the split used in prior studies. This is cited by Wu et al. in their paper on RDGCN as originating with Sun, 2018.

On page 5, there may be a typo which states "e_1 denotes another instance"; based on context, I believe that this is intended to be e_2 instead.

Reviewer #6: The paper titled "An Attribute-Enhanced Relationship-Aware Neighborhood Matching Model with Dual Attention" proposes a novel model named ARNM-DAE2A for the task of entity alignment in knowledge graphs. The goal is to match semantically corresponding entities across different knowledge graphs to facilitate knowledge fusion. The model integrates several advanced techniques, including graph convolutional networks (GCNs), dual attention mechanisms, and attribute information, to improve alignment accuracy. The authors claim that their model outperforms existing baselines on three DBP15K datasets.

The manuscript presents a well-structured and technically sound study with strong empirical validation. However, the paper suffers from several weakness.

- The manuscript lacks a rigorous theoretical discussion explaining why the dual attention mechanism and attribute-enhanced embeddings contribute to improved performance. While the empirical results demonstrate effectiveness, a deeper theoretical justification would significantly strengthen the contribution.

- The proposed model integrates multiple components, including GCN-based entity structure learning, attention mechanisms, and neighborhood-aware matching. However, the manuscript does not provide a detailed analysis of the computational cost associated with each component.

- Given the model's complexity, it is unclear how it scales to larger datasets beyond DBP15K. The authors should provide details on training time, memory usage, and computational requirements.

- While the DBP15K dataset is widely used in entity alignment research, it does not fully represent the diversity of real-world knowledge graphs. The authors should evaluate ARNM-DAE2A on additional datasets, such as YAGO, Wikidata, or Freebase, to demonstrate broader applicability.

- The manuscript lacks a discussion on the potential biases in DBP15K and whether they might influence the reported results.

- The performance of deep learning models is often sensitive to hyperparameter choices. While the manuscript provides a brief discussion on hyperparameter tuning (e.g., entity embedding dimensions, GCN layers, attention heads), it does not analyze how robust the model is to hyperparameter variations.

- Some sections, particularly in the experimental results, could benefit from more structured discussions rather than listing numerical results.

Reviewer #7: Overall, a sound manuscript describing a model that uses neighborhood matching to capture relationships between entities while applying dual attention mechanisms to weigh both the relationships between entities and their content-based features.

They have provided a thorough description of the datasets used and the statistical analysis of their approach. The figures included (especially figure 2) are helpful in illustrating the modules and relationship model framework. In addition, the authors have done a good job of stating the past research in this field and placing this work in the context of previous research. Terms are clearly defined and their approach and methodology is logically structured.

It would be helpful if the authors could further clarify the significance of their work on the impact of future applications. The manuscript would be strengthened if the authors could highlight the limitations of the advances from this proposed model and provide further details on the novelty of these new approaches in context of the current landscape.

All resources and datasets are appropriately referenced and publicly available.

The manuscript is provided in a clear and intelligible manner in standard English. There are a few minor typographical errors, including misplaced hyphenations (e.g. be-tween and in-formation) and instances of non-standard American spelling (e.g encyclopaedia). Authors should double check for typographical mistakes.

**********

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: No

Reviewer #2: Yes: Michael Winter

Reviewer #3: No

Reviewer #4: Yes: Moisés Pereira Galvão Salgado

Reviewer #5: No

Reviewer #6: No

Reviewer #7: 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.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

An Attribute-Enhanced Relationship-Aware Neighborhood Matching Model with Dual Attention

PONE-D-24-50383R1

Dear Dr. YANG,

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. If you have any questions relating to publication charges, please contact our Author Billing department directly at authorbilling@plos.org.

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,

Tao Huang

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 #2: All comments have been addressed

Reviewer #3: All comments have been addressed

Reviewer #4: All comments have been addressed

Reviewer #6: 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 #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #6: Yes

**********

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

Reviewer #2: Yes

Reviewer #3: I Don't Know

Reviewer #4: Yes

Reviewer #6: 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 #2: Yes

Reviewer #3: No

Reviewer #4: Yes

Reviewer #6: 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 #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #6: 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 #2: Thank you to the authors for addressing all comments adequately in their revision. The changes have improved the manuscript, and I appreciate their thorough response to the reviewer feedback.

Reviewer #3: I appreciate the authors for their response in addreessing my concerns. While most of my concerns have been resolved, I note that the code has not been shared due to commercial issues, and the statistical significance of the reported results remains unclear. If these limitations are acceptable, I would recommend accepting the paper.

Reviewer #4: (No Response)

Reviewer #6: The authors have adequately addressed my comments raised in a previous round of review and I feel that this manuscript is now acceptable 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 #2: Yes: Michael Winter

Reviewer #3: No

Reviewer #4: No

Reviewer #6: No

**********