How do you plan to scale content creation? Writing a book on any of the topics you plan to cover would require a lot of time, and I doubt that creating a script for a conversational agent is any easier.
Also, I'm not sure about your claim that natural language processing isn't necessary, since the dialog examples show the bot replying "correct" to a free-form answer. Or do you rely completely on the ELIZA effect making the user feel like their mistakes are being corrected when their answer differs from the explanation given by the bot?
> How do you plan to scale content creation? Writing a book on any of the topics you plan to cover would require a lot of time,
TextBooks are usually designed to serve as a reference. They tend to cover everything and has to go under a lot of scrutiny and proofreading before they can be published.
The conversational scripts can be better thought as blog posts which explains topic well but don't go into detail. You might see that the all the initial tracks are in Beta. Initially, the tracks and course content will be minimal and gradually will build up. Course Content will continuously be versioned.
> I'm not sure about your claim that natural language processing isn't necessary,
> Bot replying "correct" to a free-form answer.
That's actually is a response to multiple choice question where the user can select predefined answers.
The chatbot doesn't reply to free form answers in Boolean. To free form answers, it replys with accepted answer.
The free form answers serve two purposes:
1. To provide the user time to infer the result before Primer giving the correct answer
2. To provide the user the opportunity to verbalise his/her understanding in written form so that he/she can find out gaps between his/her knowledge and the accepted answer.
I have detailed in the whitepaper how it is possible. I can show it to with a live application.
3 comments
[ 9.7 ms ] story [ 53.8 ms ] threadIf you have any questions, please feel free to ask.
Also, I'm not sure about your claim that natural language processing isn't necessary, since the dialog examples show the bot replying "correct" to a free-form answer. Or do you rely completely on the ELIZA effect making the user feel like their mistakes are being corrected when their answer differs from the explanation given by the bot?
TextBooks are usually designed to serve as a reference. They tend to cover everything and has to go under a lot of scrutiny and proofreading before they can be published.
The conversational scripts can be better thought as blog posts which explains topic well but don't go into detail. You might see that the all the initial tracks are in Beta. Initially, the tracks and course content will be minimal and gradually will build up. Course Content will continuously be versioned.
> I'm not sure about your claim that natural language processing isn't necessary, > Bot replying "correct" to a free-form answer. That's actually is a response to multiple choice question where the user can select predefined answers.
The chatbot doesn't reply to free form answers in Boolean. To free form answers, it replys with accepted answer.
The free form answers serve two purposes: 1. To provide the user time to infer the result before Primer giving the correct answer 2. To provide the user the opportunity to verbalise his/her understanding in written form so that he/she can find out gaps between his/her knowledge and the accepted answer.
I have detailed in the whitepaper how it is possible. I can show it to with a live application.