“Standard Text” Relational Classification Model Based on Concatenated Word Vector Attention and Feature Concatenation

Round 1

Reviewer 1 Report

Review of "A Standard Text Relational Classification Model based on Cascaded Word Vector Attention and Feature Splicing"

 

The article addresses the important task of relation classification in natural language processing and highlights its significance in various NLP applications such as machine translation, human-computer dialogue, and structured text generation. The authors specifically presents a novel approach to standard text relational classification. The proposed model is based on cascaded word vector attention and feature splicing. 

Overall, this article presents a valuable contribution to the field of relation classification in standard texts. The innovative model and comprehensive experimental evaluation provide strong evidence of its efficacy. The article is well-structured, and the findings are clearly presented. 

Comments for the authors:

1. The authors refer to "standard text" and "general text" without providing a clear definition of these terms and explaining their differences. It is recommended to include a brief explanation or definition of these terms to enhance the clarity of the study.

 

2. In the introduction, the authors mention "accuracy" and measures of accuracy without explicitly explaining the differences between them. It would be helpful to provide a brief explanation or clarification of these terms to ensure a clear understanding of the evaluation metrics used in the study.

 

3. In Table 3, it is not evident why only certain F1 scores are presented for the three datasets while others are missing. It is suggested to provide a more detailed explanation of this aspect in the table caption or the results section to clarify the rationale behind the inclusion of specific F1 scores in the table.

4. Please capitalize Sections and Subsections' names.

5. It would be beneficial to provide more insights into the limitations or potential future directions of the proposed model. Additionally, discussing the practical implications and potential applications of the model in real-world scenarios would enhance the significance of the study.

Fix minor typos:

The relationship classification model based on deep learning can establish logical connections between the front and back of the text, please explain this sentence better.

Author Response

Thank you very much for taking the time to review my manuscript and highlight valuable comments. My revised response to your comments can be found in the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper titled "Standard Text Relational Classification Model Based on Concatenated Word Vector Attention and Feature Concatenation" proposes a relational classification model that utilizes concatenated word vector attention and feature stitching. The model aims to improve the performance of relation classification tasks by enhancing the positive features' effectiveness and reducing the impact of negative features. There are some observations that may be addressed for reconsideration of the revised version:

1. In the model, attention is applied in the word vector dimension, cascading the attention mechanism to amplify positive features' influence and suppress negative features' impact on relation classification. Justification with literature are missing.

2. Experimental results demonstrate that the proposed model, incorporating the two aforementioned modules, enhances the performance of relational classification. I think the limitations of this approach may be listed.

3. However, the paper acknowledges certain limitations in the proposed model. Specifically, during entity extraction, the mask matrix generation is solely based on the labels of two entities. What effects may be degraded due to this limitation need to be explained 

4. Table four may be highlighted and explained well in detail.

5. Reference 31 and 32 are duplicates; check other references for details and missing data.

6. Selection criteria for the dataset may also be highlighted by the authors

Improve the prose of the paper.

Author Response

Thank you very much for taking the time to review my manuscript and highlight valuable comments. My revised response to your comments can be found in the attachment.

Author Response File: Author Response.pdf