1. 违背了研究方向的假设;
  2. 虽然实验结果不错,但是没有明确地指向任何成功的方向,作者也没有充分地处理失败的案例——The results, though good are not clearly pointing to any direction of success, and the authors have not addressed failure cases sufficiently.(没有分析效果不好的原因);
  3. 一些小问题是论文需要一些校对,因为有些地方很难阅读——Another minor issue that the reviewer noticed is that the paper needs some proof-reading as it is hard to read in places.(英文写作加强,检查语法问题);
  4. 数据都是小数据集。这种方法在更大的数据集上有多大的实用性?在某些数据集上的计算似乎非常不切实际。——These are small datasets. How practical is this approach on larger datasets?The calculations on some of the data sets seem impractical computationally.(实验数据集太少,换大数据集跑);
  5. 请清楚的解释你是如何从数据集创建xxx方法设置的。xxx等人使用了几种不同的设置,仅仅指出这项工作是不够的,如果在你的工作中也总结了这个设置,那就更好了。—— Please explain clearly how you created the xxx setup from your datasets. xxx et al, use several different setups, it is not sufficient to merely point to that work. It would be better if the setup was summarized in your work as well.(参数分析)
  6. 定理的证明有可修改的地方和实验评价较弱——Fair: The paper has minor technical flaws. For example, the proof of a theorem has some fixable errors or the experimental evaluation is weak.(公式推导有问题、实验设计薄弱);
  7. 研究的动机是什么?是否有实际应用程序需要这样的属性——What is the motivation for studying the robustness of MIL? Are there any real-world applications requiring such a property?(研究动机
  8. 本文缺乏对所提出方法的实证结果进行深入或者深刻的观察和讨论——This paper lacks in-depth or insightful observations and discussions about the empirical results of the proposed methods.(类似问题2);
  9. 本文使用的数据集是小规模的,而且是不合时宜的。应该使用更近期和更具挑战性的数据集进行实验,以提供有价值的实验结果。——The datasets used in this paper are small-scale and out-of-time. More recent and challenging datasets should be involved as a test bed for giving valuable experimental results.(用大规模数据集和更新的数据集
  10. 所提出的方法简单,缺乏技术上的新颖性 ——The proposed methods are straighforward and lack technical novelty;


  • 能提供关键的资源,让其他研究者能复现该方法——Good: key resources (e.g., proofs, code, data) are available and sufficient details (e.g., proofs, experimental setup) are described such that an expert should be able to reproduce the main results.
  • 该论文很容易理解——The paper is easy to follow;



  1. 公式细节问题,如属于和包含于是否用对;
  2. 总体框架是不完整的——The Overall Framework is incomplete;
  3. 引号用错了——Quotation marks used wrongly;
  4. 应该增加对伪代码部分的描述——Simple description of the pseudocode section should be added;
  5. 我无法理解所提出算法的复杂性的讨论,特别是与其他算法的比较。——I could not follow the discussion on complexity of the proposed algorithms especially comparison with others;
  6. 结论部分很短,不够全面。——The conclusion section is short and not comprehensive.
  7. Xij不是向量,所以x不应该加粗。类似地,作为一个集合,B应该是粗体。——xij is not a vector, so x should not be bold. Similarly, as a set, B should be bold.(变量写法
  8. 在式(6)和式(7)中,T的值域是多少?它是Tr的子集吗?——In Equations (6) and (7), what is the range of T ? Is it a subset of Tr?(变量取值范围);
  9. 要转置一个向量或矩阵,你应该使用T而不是T——To transpose a vector or a matrix, you should use T instead of T;
  10. 文章的引言部分不仅要描述xx的整体现状。本文研究了xxx的优化算法。因此,绪论部分应重点介绍xxx的应用背景和现状。——The introduction part of the article should not only describe the overall current situation of xx. This article studies the optimization algorithm of xxx. Therefore, the introduction part should highlight the application background and current situation of xxx.(着重介绍与研究相近的工作背景)
  11. 未提供这些分类器的重要细节。 例如:J48是决策树中的一个模型。 如果这里只描述J48,是不准确的。 因此,文章中的相应信息需要更改。——it does not provide important details of these classi ers. For example: J48 is a model in the decision tree. If only J48 is described here, it is not accurate. Therefore, the corresponding information in the article needs to be changed.
  12. 本文第4部分的第二小节描述了常用的数据集。它应该突出这些数据与多实例学习的相关性,而不是简单地描述数据的大小。——The second subsection of Part 4 of the article describes commonly used data sets. It should highlight the relevance of these data to multi-instance learning, rather than simply describe the size of the data.(对数据集的相关性描述);
  13. 建议在表3、表4、表5中增加数据集特征维和样本量的描述信息,使优化算法对于不同大小数据集的通用性更加直观。——It is recommended that Table 3, Table 4, and Table 5 increase the description information of the feature dimension of the data set and the sample size, so that the versatility of the optimized algorithm for data sets of dierent sizes can be seen more intuitively.(数据集属性描述);
  14. 本文所提出的比较算法普遍已经过时,应与最新的相关方法进行比较,以显示其优越性。——The compared algorithms presented in the paper are generally out of date, it should be compared with more recent related methods to show the superiority.(用最新的对比方法);
  15. 在方法设计中,应该探讨数据集分割的必要性——Authors should discuss the necessity for the data splitting.(方法设计
  16. 在摘要和引言中,xxx被多次提及,却没有解释它的含义。——In the abstract and introduction, xxx has been mentioned many times without explaining its meaning.(未被描述);
  17. 符号被冗余定义——Symbols are defined redundantly;
  18. 写作质量需要显着提高。 在目前的形式中,手稿中有许多错字,这表明手稿没有经过仔细校对。 举几个例子:——The quality of writing needs signifi cant improvement. In its current form, there are many typo within the manuscript which suggest that the manuscript has not been carefully proofread. To name a few:Page 11, Line 52: Datesets(写作细节);
  19. 许多语法错误需要改正。举几个例子:——Many grammatical mistakes needs addressing. To name a few:Page 12, Line 35: The follow research" ! the following(语法错误);
  20. 格式化也需要工作。在参考文献中,有的期刊名称有缩写,有的没有。一些会议的参考资料列出了一个城市/地点,而另一些则没有。一些条目有DOI,而其他条目没有。所有这些形成问题都应在提交前仔细检查。——Formatting also needs work. In the references, some journal names are abbreviated while others are not. Some conference references are listed with a city/location, while others have not. Some entries have DOI, while others have not. All these formmatiing issues should be carefully checked before submission.(参考文献格式统一);
  21. 在导论中应该有一些关于xxx的讨论,并且遗漏了一些与xxx相关的参考文献。——There should be some discussion of xxx in the introduction and some xxx-related references are missed.(相关工作不足);
  22. 请在GitHub“github.com/…”的超链接前加上“https://”。——Please add “https://” before the hyper-link of GitHub “github.com/…”.
  23. 请在表III和表IV中的方法中,用粗体强调每个数据集的最佳结果。——Please emphasize the best results with the bold font for each dataset among the methods in table III and IV.(高亮最佳结果);
  24. 请重新排列图的顺序,因为图 3 就在图 8 的下方。——Please rearrange the gure order as gure 3 is right below gure 8.(图片排版);
  25. 应该客观地分析这种方法的缺点。——The disadvantages of the approach should be analyzed objectively.(分析方法缺点);
  26. 可以通过清楚地指出论文的主要贡献来改进摘要。——The abstract may be improved by clearly indicating papers main contributions.
  27. 术语应该解释清楚——The two terms xxx and xxx should be explained clearly in details.(术语要有解释

Information Science


  1. 为了可读性,本文的写作质量应该大大提高,有很多写作细节应该改进。应避免使用第一人称(即“我们”),最好使用被动语态或其他方式。——In order to readability, the writing quality of this paper should be greatly improved, there are many writing details should be improved. The use of first persons (i.e., “we”) should be avoided, and can preferably be expressed by the passive voice or other ways.(用被动语态
  2. 公式格式应该一致,请检查并修改——Formula format should be consistent. Please check and modify again;
  3. 需要改进介绍部分。作者应该更专注于他们的工作,清楚地描述具体的问题和他们的动机。需要更好地突出贡献和说明未来潜力。——The improvement for the introduction section is required. Authors should more focus on their work and clearly describe the specific problem and their motivation. Need to better highlight contributions and state future potential.(引言更加清楚描述其他工作,本文的贡献);
  4. 图中没得描述清楚文中方法,应举例详细描述;
  5. 实验部分应该提供参数设置和对结果的影响—— The experimental parameters settings and effect on the results should be provided.(参数设置合理性和分析);
  6. 在引言部分,应该有一个整体的框架,以便读者更好地理解论文的内容和方法的新颖性。——In the Introduction, there should be a holistic framework for readers to better understand the content of the paper and the novelty of the approach(开局一张图);


Pattern Recognition


< <上一篇