手把手教學如何有效回復審稿人的comments手把手教學|如何有效回復審稿人的comments每一篇論文都是科研人員的心血，投稿後“漂泊”在外（under review），“老母親”在家日夜惦念 (wait…)。終有一天，它帶著沉甸甸的“包裹”回來了 (reviewer comments)，至于“包裹”裡是驚喜還是驚嚇，包藏著一段怎樣艱辛的歷程，我們又該如何對其“梳妝打扮”甚至“改頭換面” (response to reviewers’ comments)，送他“安全”返回 (revised submission) 並抵達終點 (accept) 呢……小編帶你且看且分析。審稿人的comments千差萬別，不過依舊可以找到些許規律可循。小編認為，回復審稿人主要需要把握四點：端正個人態度、吃透反饋問題、分析回復思路、整合回復語言。Tip 1 端正個人態度一般情況下，審稿人多為高校或研究機構的科研工作者，他們或常年奮鬥在科研一線，經驗豐富；或早年“叱□”學術圈，功成身退。不管是哪種審稿人，他們也都是人，不是神。因此，存在對作者論文的內容理解不到位，所提出的問題不是非常合情合理的可能，甚至可能存在觀點偏頗；亦或審稿人對作者論文所述觀點理解極深，所提出的問題異常尖銳，

高雄市投注站

搞的作者一時之間不知如何回復。但無論什麼樣的審稿人，提出什麼樣的問題，作者首先在態度上應該做到不卑不亢、張弛有度。帶著學術交流與溝通學習的目的來審視這些問題，卸下心中的惶恐與不安，保持自信。良好的開端是成功的一半，接下來才是重頭戲。Tip 2 吃透反饋問題認真逐字逐句研讀審稿人的comments，這是做好回復的重要基礎。審稿人的問題基本上可以分為參考文獻問題、論文內容/寫作問題、論文方法問題三類。接下來我們通過舉例逐個分析。（例子涉及專業知識，請勿糾結）1、參考文獻問題具體案例：(1) Reviewer#1錛 Concern # 1: Add more refences to the literature when you talk about the pre-processing錛 feature extraction錛 machine learning錛 and deep learning methods. I suggest the following references.□ A.Alqudah錛 Ali Mohammad. “An enhanced method for real-time modelling of cardiac relatedbiosignals using Gaussian mixtures.” Journal of medical engineering & technology 41.8 (2017):600-611.□ B.Alqudah錛 Ali Mohammad錛 et al. “Developing of robust and high accurate ECG beatclassification by combining Gaussian mixtures and wavelets features.” Australasian physical &engineering sciences in medicine 42.1 (2019): 149-157. □ C.Alquran錛 H.錛 et al. “ECG classification using higher order spectral estimation and deep learning techniques.” Neural Network World 29.4 (2019): 207-219.(2) Reviewer#2錛 Concern # 2: Literature improvement is still needed. Please compare your work with some recent work錛 e.g.錛 □ J. Huang錛 B. Chen錛 B. Yao and W. He錛 “ECG Arrhythmia Classification Using STFT-Based Spectrogram and Convolutional Neural Network錛” in IEEE Access錛 vol. 7錛 pp. 92871-92880錛 2019. □ B. Hou錛 J. Yang錛 P. Wang and R. Yan錛 “LSTM-Based Auto-Encoder Model for ECG Arrhythmias Classification錛” in IEEE Transactions on Instrumentation and Measurement錛 vol. 69錛 no. 4錛 pp. 1232-1240錛 April 2020. In Sec. II.C錛 when “image classification錛 object detection錛 and image segmentation” are mentioned錛 please also cite some recent work like: □ T. D. Pham錛 K. Wardell錛 A. Eklund and G. Salerud錛 “Classification of short time series in early Parkinsonʼ s disease with deep learning of fuzzy recurrence plots錛” in IEEE/CAA Journal of Automatica Sinica錛 vol. 6錛 no. 6錛 pp. 1306-1317錛 November 2019. 2019. Y. Tian錛 X. Li錛 K. Wang and F. Wang錛 “Training and testing object detectors with virtual images錛” in IEEE/CAA Journal of Automatica Sinica錛 vol. 5錛 no. 2錛 pp. 539-546錛 Mar. 2018.案例剖析：上述審稿人的2個問題均涉及參考文獻問題，主要聚焦整篇文章的參考文獻的質量與選取，

宏碁股價

或者針對某一部分內容，如（2）中審稿人提出預處理、特征提取、機器學習和深度學習方法部分的參考文獻需要進一步優化。同時，所有審稿人均列出了可供引用的參考文獻。針對于上述涉及參考文獻類問題的審稿意見，我們首先要對審稿人推薦的參考文獻進行篩選，對于與文章內容沒有直接關系的參考文獻不予引用（但要說明原因），對于與文章內容有密切關系的參考文獻則“能引盡引”（不可否認的是，審稿人推薦的文獻很可能是本人實驗室發表的與審稿人研究領域相關的文獻，旨在增加其科研成果的曝光度）。與此同時，作者也可根據審稿人的意見對參考文獻進行合適的增刪，涉及性能比對等內容時，最好選擇近兩年的參考文獻，以保證論文的“優勢”。回復策略：三步走策略：首先是感謝審稿人的審稿工作，然後逐個分析審稿人推薦的參考文獻，並適當與本文的研究內容進行比較，保留需要參考的文獻，去除與本文無關的參考文獻並說明原因，最後闡明文中的修改位置和內容。據此，我們對上述問題（1）進行了回復以供參考。(1) Reviewer#2錛 Concern # 2: Author response:We want to thank reviewer for constructive and insightful criticism and advice about the reference. We addressed all the points raised by the reviewer as summarized below.（表示感謝）The reference □ above proposed an ECG arrhythmia classification method using two-dimensional (2D) deep convolutional neural network (CNN). Using ECG recordings from the MIT-BIH arrhythmia database as the training and testing data錛 the classification results show that the proposed 2D-CNN model can reach an averaged accuracy of 99.00%. Compare with our work錛 even though the performance metrics (accuracy錛 sensitivity錛 and specificity) of our paper are more comprehensive than the metrics (only accuracy) of □ the average accuracy from reference □ is better than our model result. And the deep learning method of STFT-Based Spectrogram provide a new idea in future work. (分析與比較)The reference □ above introduced a novel deep learning approach that integrates a long short-term memory (LSTM)-based auto-encoder (AE) network with support vector machine (SVM) for electrocardiogram (ECG) arrhythmias classification. The proposed method achieved average accuracy錛 sensitivity錛 and specificity of 99.45%錛 98.63%錛 and 99.66%錛 respectively錛 in the beat-based cross-validation approach based on the Advancement of Medical Instrumentation (AAMI) standards. The performance is superior to our model result (average accuracy錛 sensitivity錛 and specificity of 98.00%錛 96.17%錛 and 96.38%). We think the LSTM-based auto-encoder (AE) network is a positive strategy錛 which can effectively extract the characteristic information of time series signals. (分析與比較)Hence錛 we supplement some recent work in Table 10. The comparative literature is comprehensive錛 which not only include literature of the machine learning method and deep learning method錛 but also the main work is the two or multi-classification of arrhythmia signal. And most literature also is recent years (2019 and 2020). (修改內容)Meanwhile錛 we also add the recent literature like □ and □ above in Sec. II.C. when “image classification錛 object detection錛 and image segmentation” are mentioned. The classic literature and latest literature complement each other to make the reference literature more substantial and reasonable. (修改內容)2、論文內容/寫作問題具體案例：(1) Reviewer#3錛 Concern # 2: In Section C. Machining Learning Methods錛 the authors mention many disadvantages of the reviewed methods錛 such as poor robustness錛 overfitting and bad performance in practical applications. Since many shortcomings have been mentioned錛 the authors are supposed to find a way to solve them. However錛 in the experimental part錛 it is not clear whether the proposed method can overcome the shortcomings mentioned in Page 4 (主要問題1). Only results with high performance metrics are presented. It makes the focus of this paper unclear (主要問題2). In my view錛 these shortcomings mentioned in Page 4 are common among existing machine learning methods. The authors are encouraged to dig deeper and conclude the disadvantages and challenges of existing methods for ECG signal detection and classification錛 and then propose your method and highlight its difference and superiority (修改意見1). In my view錛 the major contribution of this paper is the new design of a network architecture. The authors are encouraged to highlight this and compare with other related and similar networks (修改意見2).(2) Reviewer#1錛 Concern # 1:I do remember that I have reviewed this paper錛 and the current version is an improved version of that paper. I have also seen that the authors did a good job in improving the quality of paper錛 although I personally do not like the organization and format of the paper.It also seems that the proposed method could not outperform the best results from the literature (主要問題1). But in my opinion錛 it is still worthwhile publishing this paper to let others know that this approach has already been tested and compared to other methods could achieve a relatively less accurate result.Finally錛 I suggest that the authors improve the quality of their paper and avoid English problems (主要問題2)錛 and if possible錛 shorten the paper錛 as it is quite lengthy (主要問題3).案例剖析：上述審稿人的兩個問題均涉及論文內容/寫作問題，而且兩個問題比較長，所以要分析提取出主要問題。問題（1）中涉及兩個主要問題，主要問題（1）是實驗部分沒有清楚表明論文所提模型解決了Section C中文獻綜述所闡述的目前方法的缺陷；主要問題（2）是論文模型僅展示了非常好的性能，但是論文的焦點不是很清晰。這兩個主要問題其實表述較為模糊，很難找到很好的切入口進行回復和並作出針對性的修改。但是，審稿人同時也給出兩點修改意見。所以，針對這類問題，我們只需根據審稿人給出的兩個修改方向進行修改即可。問題（2）中涉及了3個主要問題：論文模型性能並沒有優于對比文獻的所有參考文獻（雖然審稿人表明模型性能不是最優，但論文仍有價值。我們依然要對這個問題進行回復）、論文英語語法需要修正和論文內容需刪減。回復策略：依然三步走策略：首先是感謝審稿人的審稿工作，然後分析審稿人comments中的主要問題和修正意見，最後針對審稿人所提出的主要問題進行逐條回復並修正。具體回復：(1) Reviewer#3錛 Concern # 2: Author response:Thank you very much for your careful review and constructive suggestions with regard to our manuscript. Meanwhile錛 I am also very grateful to you for the specific modification strategy. (表示感謝) I have made comprehensive and detailed changes according to your suggestions. On one hand錛 we dig deeper and conclude the disadvantages and challenges of existing methods in section II. On the other hand錛 we also highlight the proposed model’s difference and superiority and compare with other related and similar networks. Below is our point-by-point modification to the referee’s suggestions.(1) The disadvantages and challenges of existing methods for ECG signal detection and classification. (問題剖析)Broadly speaking錛 the fundamental disadvantages and challenges of existing machine learning methods for ECG signal detection and classification are that hand-crafted extracted feature not only greatly affects the accuracy of the algorithm錛 but also consumes a lot of calculation time and cost. The deep convolutional neural network is essentially realized by stacking automatic encoders. Considerable feature representational power effectively reveals unknown abstract features of input signals. It can achieve self-learning through end-to-end model design. Meanwhile錛 the radical problem of both methods is that they only focus on how to propose a better model錛 but do not pay attention to data processing issues: such as data denoising錛 data augmentation錛 and multi-scale data training and testing. The data preprocessing of signals should be focus on because signals and images are different data types.Hence錛 in this work錛 a more accurate and robust method based on deep learning is proposed to identify five different types of arrhythmia signals. The proposed model not only pays attention to the superiority of model design but also presents the importance of data processing in this paper. The final results also prove that the application of ECG signal classification using the convolutional neural network is reliable. The deep learning architecture outperforms the hand-crafted feature extractors assembled by machine learning models in terms of classification accuracy錛 sensitivity錛 specificity錛 and confusion matrix.(2) The contributions of this work and model’s superiority are as follows: (問題剖析)□ We propose an end-to-end plain-CNN architecture and two MSF-CNN architectures (A and B) to replace additional hand-crafted feature extraction錛 selection錛 and classification using machine learning methods. The plain-CNN is a baseline model錛 the MSF-CNN A and B are implemented on this baseline. Thus錛 it significantly enhances the performance against recent state-of-the-art studies. The best model MSF-CNN architecture B achieves an average accuracy錛 sensitivity錛 and specificity of 98.00%錛 96.17%錛 and 96.38%錛 respectively. This illustrates the method with residual learning and group convolution blocks has a profound effect on the feature learning of the model. The results of ablation experiments show that our proposed biometric recognition and diagnosis network with residual learning (MSF-CNN B) achieves a rapid and reliable diagnosis approach on ECG signal classification錛 which has the potential for introduction into clinical practice as an excellent tool for aiding cardiologists in reading ECG heartbeat signals. (分層回答)□ Moreover錛 the signal processing problems are fully considered. We first design multi-scale input signals錛 including 251 samples (named set A) and 361 samples (named set B). This design can improve the generalization ability of the model by extracting multi-scale signal features. Then錛 the signal denoising and data augmentation also are implemented in this paper. (分層回答)□ In addition錛 we present six sets of detailed ablation experiments on ECG signal classification and achieve excellent performance metrics. Our residual learning network with group convolution (including parallel and concatenation group convolutional block) can easily achieve accuracy gains from greatly increased depth錛 producing results substantially better than the plain-CNN and MSF-CNN A. It also demonstrates deep learning can be leveraged as feature learning mechanisms to classify biomedical signals. (分層回答)□ In particular錛 we compare the results from our model to recent state-of-the-art methods. Additionally錛 detailed analysis and comparison are presented in this paper. (分層回答)Finally錛 in this manuscript錛 we replace some old references with the recent methods published in 2019 and 2020 in Table 10. Meanwhile錛 we have also done a full comparison and analysis.(2) Reviewer#1錛 Concern # 1:Author response:We would like to thank the reviewer for their thoughtful review of our manuscript (表示感謝). According to the reviewer’s comments錛 I think that the reviewer has no serious problems about this manuscript. There are three problems that require minor edits. The first minor problem is that the proposed method could not outperform the best results from the literature. The second minor problem is that the English expression needs to be improve. And the last minor problem is that the paper needs to be shorten if possible.(1) The first minor problem is that the proposed method could not outperform the best results from the literature. (問題剖析)In [62]錛 the performance is superior to our models’ result (average accuracy錛 sensitivity錛 and specificity of 98.00%錛 96.17%錛 and 96.38%). We think the LSTM-based auto-encoder (AE) network [62] is a positive strategy錛 which can effectively extract the characteristic information of time series signals. Most importantly錛 the method in [62] is the key point worthy of our study in future work. We will fully consider the optimization methods of [62] in our future work.Compared with [62]錛 our research mainly has the following differences:First錛 compared with the LSTM-based auto-encoder network in [62]錛 our model is more lightweight and less computationally expensive. The LSTM is a bidirectional model錛 which is utilized to extract the bidirectional information from the forward model (operating from t0錛 t1錛 …錛 to tn.) and backward model (operating from tn錛 tn-1錛 …錛 to t0.) at the same time. There is no doubt that the advantage will also cost a lot of computational expensive. (分層回答)Second錛 in recent years錛 the CNN (Convolutional Neural Network) and RNN (Recurrent Neural Network) is two popular deep learning methods to process the time series data. Both methods have their own advantages. The LSTM is a recurrent neural network that has an input layer錛 two hidden layers錛 and an output layer. The hidden layers are a memory block included three gate mechanisms called input gate錛 forget gate錛 and output gate. These gate mechanisms control the amount of information and the extraction of features. (分層回答)In additional錛 [62] is also the latest reference published in 2020 that I added in the last revision. The reason why I cite this paper with better performance index is we think the auto-encode network based on LSTM is a positive idea. Our paper has gone through a long review process錛 during which the associate editor and reviewer have been changed. Compared with [62]錛 the performance advantage is really not obvious. Hence錛 we will try to propose a lightweight model based on LSTM to deal with this problem in future work. (分層回答)(2) The second minor problem is that the English expression needs to be improve. (問題剖析)We tried our best to improve the expression of grammar錛 typos and professional representation of this manuscript and made many changes in the manuscript. And we also leverage 3rd party service for language polishing. we take great care to revise the grammars錛 typos and professional representation of our manuscript. (3) And the last minor problem is that the paper needs to be shorten if possible. (問題剖析)This paper has been reviewed by many reviewers錛 and a lot of contents have been added to the original text according to the requirements of the reviewers. Therefore錛 many revisions of this paper lead to the final paper being too long. In this manuscript錛 we also deleted some content appropriately as soon as possible. In section II. Related work錛 we deleted the detailed description about the references’ contents on the basis of retaining relevant research references. In section III and VI錛 we also deleted inessential contents. （修改內容）3、論文方法問題具體案例：(1) AE’s comment錛 Concern # 1:Also錛 finding optimal parameters to ensure the network performance is important by using advanced intelligent optimization methods錛 e.g.錛□ J. J. Wang and T. Kumbasar錛 “Parameter optimization of interval Type-2 fuzzy neural networks based on PSO and BBBC methods錛” IEEE/CAA J. Autom. Sinica錛 vol. 6錛 no. 1錛 pp. 247−257錛 Jan. 2019 and S. Gao錛 et al.錛 □ “Dendritic neuron model with effective learning algorithms for classification錛 approximation and prediction錛” IEEE Transactions on Neural Networks and Learning Systems錛 30(2)錛 pp. 601 – 614錛 Feb. 2019. I would suggest the authors to treat it as future work.(2) Reviewer#2錛 Concern # 3: In [63] and few others錛 the accuracy is better in others’ work. The authors need to discuss where their work stands with other works on 5-class classification problem.案例剖析：上述問題（1）中涉及模型參數優化問題，並給出具體的模型參數優化的參考文獻。面對這樣的問題往往很迷茫，如果重新選擇參數優化方法，那就面臨要再次做具體的實驗，並且和本文的實驗進行分析比較。小編感覺這樣修正內容過多，而且可能會偏離本論文所論述的具體問題，很沒有必要再增加實驗，其實審稿人提出的這個問題小編個人感覺重點不是很突出，推薦將這一問題轉換成參考文獻的問題。首先依然是闡述並篩選出有效的參考文獻，然後分析參考文獻的方法與本文的方法的異同點，重點突出本文的優勢和審稿人推薦文獻所用方法的可取之處。上述問題（2）也非常具有代表性，主要表述作者提出的參考文獻[63]的性能要優于文章模型的性能，可謂是作者自己給自己挖了個大坑。但是，作者原本意圖是想闡述文獻[63]中有些優勢可以借鑑，是以後研究的重要方向，客觀的來講，每篇文章不可能盡善盡美，除了優勢之外，對于limitation的闡述也是非常重要的。面對這一問題，主要策略首先還是作者需要闡明本意，然後分析比較兩篇文章，突出本文優勢。回復策略：重要的“三步走策略”說三遍：首先是感謝審稿人的審稿工作，然後深入剖析審稿人comments中的主要問題，不用遺漏審稿人提出的所有細節問題，最後針對審稿人所提出的主要問題進行逐條回復並修正，重點要突出本論文方法設計的主要優勢。具體回復：(1) AE’s comment錛 Concern # 1:Author response:Thanks a lot for the AE’s comments about the parameter optimization. In this paper錛 we adopt some effective parameter optimization methods. （表示感謝）First錛 back-propagation algorithm. In this paper錛 we proposed three convolutional neural networks. The parameters update of the models mainly rely on the back-propagation algorithm. The gradient is backed to the shallow layers of network so that implement the update of weight and bias. (分層回答)In addition錛 in the experiment錛 the L2 norm of the model parameters (equation (1)) is implemented to optimize gradient and relieve overfitting. Specifically錛 the threshold is set to 0.5 to stabilize the training process.where is the loss function with L2 regularization and is the cross-entropy loss function from equation (9). denotes a penalty factor錛 which is to balance the goal of achieving better training results and keeping smaller parameter values. Thus錛 the regularization can avoid overfitting effectively by narrowing down all the parameters. describes the weight of layers. (分層回答)As mentioned in reference □ above錛 in this paper錛 interval type-2 fuzzy neural networks (IT2FNNs) is the hybridization of interval type-2 fuzzy systems (IT2FSs) and neural networks (NNs). Big bang-big crunch (BBBC) optimization and particle swarm optimization (PSO) are applied in the parameter optimization for Takagi-Sugeno-Kang (TSK) type IT2 FNNs. The employment of the BBBC and PSO strategies can eliminate the need of backpropagation computation. The computing problem is converted to a simple feed forward IT2 FNNs learning. The adoption of the BBBC or the PSO will not only simplify the design of the IT2 FNNs but will also increase identification accuracy when compared with present methods. (分析與比較)In reference □ above錛 in this paper錛 these problems錛 such asthe hard understanding problem錛 the slow and difficult training problems錛 and the difficulty to scale them up錛 motivate us to develop a new dendritic neuron model (DNM) by considering the nonlinearity of synapses錛 not only for a better understanding of a biological neuronal system錛 but also for providing a more useful method for solving practical problems. To achieve its better performance for solving problems錛 six learning algorithms including biogeography-based optimization錛 particle swarm optimization錛 genetic algorithm錛 ant colony optimization錛 evolutionary strategy錛 and population-based incremental learning are for the first time used to train it. The best combination of its user-defined parameters has been systemically investigated by using the Taguchi’s experimental design method. (分析與比較)In the future錛 we will implement the parameter optimization methods mentioned above to achieve its better performance.At last錛 in this manuscript錛 we found many grammatical and representational errors in the paper. Hence錛 we invite a 3rd party service for language polishing before resubmitting. The grammar and expression have been greatly improved in this paper. We would like to submit the enclosed manuscript錛 which we wish to be considered for publication in IEEE Access. I carefully revised all the questions given by two reviewers. Meanwhile錛 we have made detailed response in the supplemental files. (分層回答)(2) Reviewer#2錛 Concern # 3: Author response:We would like to thank the reviewer for your thoughtful review of our manuscript. Indeed錛 in [62] (The [63] in last manuscript is renamed as [62] in this manuscript)錛 the performance (average accuracy錛 sensitivity錛 and specificity of 99.45%錛 98.63%錛 and 99.66%) is superior to our models’ result (average accuracy錛 sensitivity錛 and specificity of 98.00%錛 96.17%錛 and 96.38%). We think the LSTM-based auto-encoder (AE) network used in [62] is a positive strategy錛 which can effectively extract the characteristic information of time series signals. We will fully consider the optimization methods of this paper [62] in our future work. Compared with [62]錛 our research mainly has the following differences:First錛 compared with the LSTM-based auto-encoder network in [62]錛 our model is more lightweight and less computationally expensive. The LSTM is a bidirectional model錛 which is utilized to extract the bidirectional information from the forward model (operating from t0錛 t1錛 …錛 to tn.) and backward model (operating from tn錛 tn-1錛 …錛 to t0.) at the same time. There is no doubt that the advantage will also cost a lot of computational expensive. (分層回答)Second錛 in recent years錛 the CNN (Convolutional Neural Network) and RNN (Recurrent Neural Network錛 LSTM is a replacement of the traditional RNN) is two popular deep learning methods to process the time series data. Both methods have their own advantages. The LSTM is a recurrent neural network that has an input layer錛 two hidden layers錛 and an output layer. The hidden layers are a memory block included three gate mechanisms called input gate錛 forget gate錛 and output gate. These gate mechanisms control the amount of information and the extraction of features. (分層回答)Third錛 it also seems that the proposed method could not outperform the best results from the literature. But in my opinion錛 compared with most of performance results in Table [9]錛 the results of our paper is advantageous錛 and it is still worthwhile publishing this paper to let others know that this approach has already been tested and compared to other methods could achieve a relatively less accurate result. Most importantly錛 the method in [62] is the key point worthy of our study in future work. (分層回答)In additional錛 [62] is also the latest reference published in 2020 that I added in the last revision. The reason why I cite this paper with better performance index is we think the auto-encode network based on LSTM is a positive idea. Our paper has gone through a long review process錛 during which the associate editor and reviewer have been changed. Compared with [62]錛 the performance advantage is really not obvious. Hence錛 we will try to propose a lightweight model based on LSTM to deal with this problem in future work. (分層回答)Tip 3 分析回復思路小編認為“思路”是回復審稿人問題的重中之重，主線明確、思路清晰，能夠使得審稿人讀起來清楚明了，這一點十分重要。看過上面Tip2中對于具體問題的剖析，

註冊送點數

相信你已經對小編推薦的回復思路有了大概的了解。首先，無論審稿人給出的問題尖銳與否、合理與否、正確與否，我們都要對其表達謝意，這是對審稿人工作的必要性尊重；其次，分析審稿人的comments，提取所有涉及到的問題，並分段、分層將問題進行描述；第三，針對每個問題進行回復，如果問題需要大量重點闡述，那就需要作者分層、分點、遞推式的闡述；最後需要適當描述對于審稿人comments的具體修改方式和修改位置（如第幾頁第幾行）。Tip 4整合回復語言對于審稿人的comments逐條回復，突出回復重點、要點以及希望引起審稿人注意的點是整合回復語言的主要工作。另外，回復問題之前，對審稿人的審稿工作表示感謝也是不多餘的，畢竟給審稿人留個懂禮貌知禮節的好印象或許有助于審稿人“開開心心”工作。因此，小編通常推薦以下回復模式：(1) Reviewer#1錛 Concern # 2: 列出每位審稿人的每條commentsIn Table 10錛 the authors compare their work with others’ work. But do others’ work have same experiment conditions錛 e.g.錛 dataset and classes?(2) Author response: 對審稿人的comments進行個人陳述We want to thank reviewer for constructive and insightful criticism and advice about the Table 9 (Table 4 is deleted in this manuscript錛 and the Table 10 is named as Table 9) (表示感謝). We addressed all the points raised by the reviewer as summarized below.（分點闡述）First錛 we readjusted the contents of Table 9 and supplemented the column of database and main work. In Table 9錛 we listed the 12 related papers including publication time錛 main work錛 database錛 main approach錛 and performances (accuracy錛 sensitivity錛 specificity). These papers were published in 2019 and 2020 and represented the latest research results. And the main work is the classification of arrhythmia signals. It's just that the number of categories in some papers is different. The used database mainly includes MIT-BIH arrhythmia database and Chinese Cardiovascular Disease Database錛 of which most studies use MIT-BIH arrhythmia database. This is consistent with our research paper. The main methods used in this comparative literature are machine learning and deep learning錛 which is mainly to compare the final performance of the two methods.Second錛 we also fully analyzed the similarities and differences between the research from others’ work listed in Table 9 and our research in this paper in section V. C. (performance comparison and discussion). The main research work of these references listed in Table 9 is to detect and classify arrhythmia signals. Almost all studies use the MIT-BIH arrhythmia database except reference [53]錛 [56] and [60]. However錛 the different databases have little impact on the final analysis. Because the original data from two databases need to pre-processing錛 such as wave detection錛 normalization錛 signal segmentation錛 and signal annotation (This work was completed by experts of the database publishers in the early stage). In additional錛 we mainly summarized the research literature on the use of machine learning and deep learning methods錛 which is to analyze the advantages of deep learning method. (3) Author action: 簡要說明具體的修改方法與具體位置We firstly updated the manuscript by readjusting the contents in Table 9 and supplementing the column of database and main work in Table 9. In additional錛 we also fully analyzed the similarities and differences between the research from others’ work listed in Table 9 and our research in section V. C. All the corrections were highlighted using a yellow highlight tool within the new pdf file.因此，小編推薦的回復模式主要由以下三個部分構成：羅列審稿人的comments（必要時可將其分條描述）；在回復時首先表示感謝，再通過分段、分點、分層的方式闡述觀點、修改的內容以及相應的結論；對修改的方法和位置進行簡要描述。此外，可以通過改變字體顏色、粗細程度來表示重點信息。經過一系列的“猛虎操作”，相信你已經可以基本了解審稿意見的回復過程啦。不過具體情況具體分析，不同的研究領域其學術習慣也不盡相同，因此小編鄭重提示：以上交流僅供參考，希望大家多多批評指正！象牙之塔與人間情懷武夷山（發表于《科技日報》2005年2月20日） 有些人認為，

宏碁股價

科學家，尤其是從事理論研究的科學家，乃是科學界的精英人物，

宏碁股價

其研究內容屬于陽春白雪，他們就應該呆在象牙之塔裡，不受幹擾地專注于理論探討。而從事應用開發的科技人員沒有前述科學家檔次高，其研究內容屬于下裡巴人，但是，他們具有人間情懷。二者真是這樣截然對立嗎？事實勝于雄辯。在絕大多數人心目中，

5pk玩法

達爾文這位進化論思想的先驅者，是純而又純的科學家，他與科普大概是八桿子打不著的吧。且慢。在19世紀80年代，達爾文寫了一本關于堆肥生物學的普及著作，大受歡迎。該書面市頭8個星期就銷售出了3500本。在達爾文那個時代，科學界認為，科學家為非專家撰寫普及讀物是理所當然的份內之事。﹝1﹞ 愛因斯坦更是不折不扣的純理論科學家了。沒錯。但是，您知道嗎，他1928年曾經與匈牙利血統的物理學家利奧. 西拉德（後來，愛因斯坦就原子彈研制問題的迫切性上書美國總統羅斯福，那封信其實是西拉德起草的）在英國共同申請並獲得了一份電冰箱技術的專利。其方案是，用一種金屬微粒的懸浮液來代替以往冰箱中用的有毒的冷卻劑，用一個磁泵來驅動。一家名叫AEG的德國公司根據這項專利進行了冰箱原型試制錛 可是錛 這種原型電冰箱的噪聲太大。愛因斯坦與西拉德面對這一問題，並不氣餒，從1928年年底到1930年，他倆一共登記了29項與冰箱或磁泵有關的專利。不過，後來沒有公司再對他倆的專利發生興趣。與曼哈頓工程同時，恩裡科. 費米和西拉德等科學家也開始了和平利用原子能的研究.他們試制的第一個核反應堆曾遇到堆芯散熱不佳的難題。此時，西拉德突然想起了AEG試制的冰箱原型中的磁泵，它恰好可作為堆芯的理想散熱泵。通過設計上的改進，

宏碁股價

再加上採用了更先進的電機，後來的散熱泵的噪聲也比原來磁泵的噪聲小多了。如今，世界上許多核反應堆堆芯中用以輸運液態鈉的磁泵，都是從愛因斯坦和西拉德設計的磁泵演化而來的。﹝2﹞ 參考文獻﹝1﹞dspace.dial.pipex.com/sbs/isolateion.htm錛 1999年檢索﹝2﹞武夷山，愛因斯坦、西拉德和與冰箱專利，科技日報，1990，4，18，