ICACT20210125 Question.1
Questioner: contact@technologyland.co.th 2021-02-09 오후 7:44:46 |
ICACT20210125 Answer.1
Answer by Auhor jianms@nfu.edu.tw 2021-02-09 오후 7:44:46
Chrome  Click!! |
| How long does it take for one execution? |
that's for your question.
in fact, it depends on the content of the false news and the hardware for training.
generally speaking, the execution time for the false news with multiple sources can be only less than a day.
|
ICACT20210125 Question.2
Questioner: jianms@nfu.edu.tw 2021-02-07 |
ICACT20210125 Answer.2
Answer by Auhor jianms@nfu.edu.tw 2021-02-07
Chrome Click!! |
| Is the proposed system available for different languages? the vocabulary merging is correct or not? |
In some cases, especially specific names such as locations, company names, personal names, even some new created vocabularies are not easy to identified. If the vocabulary is new created or rarely used, to find the suitable length of the merged vocabulary would cost more time. However, if the new created vocabulary now is popular and generally used, the proposed system still could find it. |
ICACT20210125 Question.3
Questioner: jianms@nfu.edu.tw 2021-02-07 |
ICACT20210125 Answer.3
Answer by Auhor jianms@nfu.edu.tw 2021-02-07
Chrome Click!! |
| How about the performance of the proposed system? How much time is needed for predicting the false news? |
Currently, we would say that the proposed system could provide the possibility of the false news according to the past data. Hence, without enough data, most news can not be predicted as the false news. Furthermore, the false news are not always the same content or broadcast by the same groups. However, to divide the news into the corresponding groups is still work. |
ICACT20210125 Question.4
Questioner: oumaymamejrix@ewhain.com 2021-02-09 오후 8:16:14 |
ICACT20210125 Answer.4
Answer by Auhor jianms@nfu.edu.tw 2021-02-09 오후 8:16:14
Chrome Click!! |
| Hello, First I would like to thank you for your work, it looks impressive, it says at the end that it is possible to process different languages because as you know there are languages that use a totally different alphabet, so can the same results be obtained? |
thanks for your question.
currently, English or most languages used in Europe can be identified since the vocabulary is break by one space. If the languages or vocabularies are not break based on the space, such as Chinese, Japanese, Korean, the corresponding break method is needed. some vocabularies should be recorded and added in the database for sentence segment breaking. however, we believe that after finding the segments from the original content, the training procedure still works. |
ICACT20210125 Question.5
Questioner: satyabrataaich@gmail.com 2021-02-09 오후 8:10:01 |
ICACT20210125 Answer.5
Answer by Auhor jianms@nfu.edu.tw 2021-02-09 오후 8:10:01
Chrome Click!! |
| Thank you for your presentation. What are the advantages of the proposed method compared to the previous methods. what is the effectiveness of the proposed approach if the morphology is different than the current one? |
thanks for your question.
the proposed system focuses on the related feature finding.
to prove the news be true or false still depends on the related proof center.
by collecting the false news, the malicious group which create the false news or broadcast false news may be found. in addition, the topic or object could be also found due to the false news features.
since the false news is broadcast by social media or online platform, when the news is proved as false news, the feature still could be analyzed. |
ICACT20210125 Question.6
Questioner: linxiaocer@fzu.edu.cn 2021-02-10 오후 5:55:48 |
ICACT20210125 Answer.6
Answer by Auhor jianms@nfu.edu.tw 2021-02-10 오후 5:55:48
Chrome Click!! |
| The YouTube video of the presentation is absent. |
Dear sir,
would you please indicate which page youtube presentation is absent?
Through the website of ICACT, the youtube links are checked and still work.
Each page is presented independently.
Please feel free to contact me if there is still any problem. |
