ICACT20210164 Question.1
Questioner: h.warunyu@gmail.com 2021-02-09 오전 4:57:00 |
ICACT20210164 Answer.1
Answer by Auhor freemanpolys@gmail.com 2021-02-09 오전 4:57:00
Chrome  Click!! |
| Can your method be used in same kind of situation except COVID-19 in the future? |
Yes. Our proposal main goal is to help the student improve themselves by freely interact with a chatbot.
Students are usually afraid to ask questions so as not to be judged. With a converstional bot, they no longer have this brake.
The teachers after their lessons do not know if the students really assimilated their lessons. With the bot, they will have a good appreciation of the mastery of their course by the students |
ICACT20210164 Question.2
Questioner: patcharin.ka@kmitl.ac.th 2021-02-09 오후 3:34:22 |
ICACT20210164 Answer.2
Answer by Auhor freemanpolys@gmail.com 2021-02-09 오후 3:34:22
Chrome Click!! |
| can your method apply with other LMS platform? and what specification that LMS platform should have for working together with NLP?(API,...???) |
Yes, the chatbot can be plugged into other LMS platform. The NLP part which hold the business logic is external to the LMS and the integration can be done via Rest API or SocketIO interface. |
ICACT20210164 Question.3
Questioner: patcharin.ka@kmitl.ac.th 2021-02-09 오후 3:49:41 |
ICACT20210164 Answer.3
Answer by Auhor freemanpolys@gmail.com 2021-02-09 오후 3:49:41
Chrome Click!! |
| why you select Rasa framework in this research? |
We select Rasa Framework because the goal of our work is to focus on how to help student in this covid-19 outbreak. Rasa Framework already have the NLP engine which can be trained quickly. It can be integrated to other platform like Facebook Messenger, Twilio (via sms, whatsapp), Telegram and custom integration is done via Rest APi or SocketIO. It was easy to integrate our NLP bot into any LMS (we integrate it into moodle by coding a plugin). As our perpective is to also remediate new students coming from another university while onboarding them, we have the choice with Rasa to evaluate this new students via Facebook or Whatsapp because they can't access the school learning platforms (id, credentials ...). |
ICACT20210164 Question.4
Questioner: alihussain.dream@gmail.com 2021-02-09 오후 5:14:28 |
ICACT20210164 Answer.4
Answer by Auhor freemanpolys@gmail.com 2021-02-09 오후 5:14:28
Chrome Click!! |
| Can you explain about features of the chatbox (like how many users can use it), and if any user wants to delete any msg or conversation is it possible? |
Honestly, we don't make a stress test yet. The chabot has 3 main components : the NLU Engine, the action server which hold the business logic and the database. We have just deployed the three component in one server and it works great with our LMS platform since all the students are not using it at the same time. We will next monitor the needs of compute and and make some stress tests.
For the second question, the chabot goal is to evaluate the students so there is not a feature for deleting messages or conversation because students mainly answer courses questions. They only have 3 chances to retake an evaluation. |
ICACT20210164 Question.5
Questioner: quandh@vnu.edu.vn 2021-02-09 오후 7:46:47 |
ICACT20210164 Answer.5
Answer by Auhor freemanpolys@gmail.com 2021-02-09 오후 7:46:47
Chrome Click!! |
| Thanks for your presentation. I'd like to know more about your training data. Did you build it from scratch or use any publicly available one.
Besides, can you please detail your parameters in order to customize the model and finetune it on your dataset? |
We build it from scratch but note that for the NLU pipeline (tokenizers, featurizers and training algorithms) we use public models : WhitespaceTokenizer, RegexFeaturizer, CountVectorsFeaturizer, DIETClassifier ... For the model, it's the DIETClassifier which is used with epochs value 100. Our configurations is based on the Rasa doc suggested one . You can find them here : https://rasa.com/docs/rasa/model-configuration |
