ICACT20220329 Question.1
Questioner: tomayoon@ieee.org 2022-02-15 오후 12:43:05 |
ICACT20220329 Answer.1
Answer by Auhor anquanip@cnic.cn 2022-02-15 오후 12:43:05
Chrome  Click!! |
| First of all, thank you for the good research paper and presentation contents. I would like to know more about the research environment used in this study, and also in detail about the equipment & operating systems, development languages and software libraries you used. |
Thank you for your question.The experimental environment is Intel (R) Core (TM) i7-7700 CPU on the Windows 10 platform, the main frequency is 3.60GHz, and the memory is 16 GB RAM,and we use python development language and scikit-learn model library. |
ICACT20220329 Question.2
Questioner: tomayoon@ieee.org 2022-02-15 오후 12:48:38 |
ICACT20220329 Answer.2
Answer by Auhor anquanip@cnic.cn 2022-02-15 오후 12:48:38
Chrome Click!! |
| Will you tell us what is the major achievement of this paper, and also what is further study issue from this? Thanks in advance!! |
This paper proposed a lightweight intrusion detection model based on feature selection and autoencoder. The model consists of the following three parts: data preprocessing, feature selection and anomaly detection. In the feature selection module, we use a recursive feature addition algorithm, which can select the optimal subset of features and relief the overfitting problem. We use One-Class SVM on the NSL-KDD to verify its effectiveness. In the anomaly detection module, we use the autoencoder, which can adapt to the problem of sample imbalance, so it is suitable for deployment in real network scenarios. Experimental results show that our model can speed up training and detection time, and accurately detect attacks. In the future research, we attempt to use real network traffic data to further evaluate our model. |
ICACT20220329 Question.3
Questioner: namacabale@gmail.com 2022-02-16 오전 11:13:20 |
ICACT20220329 Answer.3
Answer by Auhor anquanip@cnic.cn 2022-02-16 오전 11:13:20
Chrome Click!! |
| Dear Author, you have a very good research. I would like to know if the technology today is capable of implementing your proposed scheme? Is there an existing technology almost similar to your proposal that has been deployed in real world environment? Thank you and stay safe.
|
Thank you for your question.Our model can be deployed on simple network devices and is expected to realize plug and play in the future. Our model can realize real-time detection. As far as I know, no similar model has been used in the actual environment. We hope our model can spark an interest in further developing the domain of online neural network based anomaly detection. |
ICACT20220329 Question.4
Questioner: vasaka.vis@mahidol.ac.th 2022-02-16 오전 11:02:27 |
ICACT20220329 Answer.4
Answer by Auhor anquanip@cnic.cn 2022-02-16 오전 11:02:27
Chrome Click!! |
| Why you use "recall" as the performance metric in this paper? Not the accuracy, precision or F1-score? |
Thank you for your question.As we all know, in anomaly detection, people pay more attention to the recall rate(recall=TP/(TP+FN)) and even want to ensure that it is 100%.In this paper, we give the confusion matrix, so the accuracy, precision, F1 and other indicators can be calculated. We focus on the analysis of the recall and compare it with other methods, in order to emphasize the advantages of our model in anomaly detection methods. |
ICACT20220329 Question.5
Questioner: 20180922@sungshin.ac.kr 2022-02-16 오전 10:40:35 |
ICACT20220329 Answer.5
Answer by Auhor anquanip@cnic.cn 2022-02-16 오전 10:40:35
Chrome Click!! |
| Thank you for the good presentation. The data sets used in this study are NSL-KDD and CSE-CIC-IDS2018, why did you select that data set among many other data sets? |
Thank you for your question.We have compared various data sets. Nsl-kdd data set is a classic intrusion detection data set. Many studies have used it. We choose to use it for comparison with other studies. CSE-CIC-IDS2018 data set was recently generated (2018). It contains seven common update attack families that meet real-world standards and can be publicly available. The data set has the characteristics of large scale, sufficient features and many kinds,which meets the needs of our research. |
ICACT20220329 Question.6
Questioner: ycc.nttu@gmail.com 2022-02-16 오전 10:44:56 |
ICACT20220329 Answer.6
Answer by Auhor anquanip@cnic.cn 2022-02-16 오전 10:44:56
Chrome Click!! |
| Thank you for your analysis and discussion of Network Intrusion Detection. Do you have any suggestions for future applications with your way? |
Thank you for your question.Experimental results show that our model can speed up training and detection time, and accurately detect attacks. In the future research, we attempt to use real network traffic data to further evaluate our model, and further adjust the model according to the results. |
