2021 #VWEDU: #DTA Model for a chatbot working from an Expertise Database. A possible library function?

Model for a chatbot working from an Expertise Database. 

A possible library function?

Libraries answer questions.  FAQ pages answer questions. An online library could automate that.

We are not alone

We digitals are not alone

• Not even in quarantine
• Not even if the people nearby don't share our interests.
• In a digital world we find people who share our interests.
• People who are doing the kinds of things we do.
• People who have questions we can answer.
• People who have answers to some of our questions.
• People who hold meetings about things we are interested in.
• Era of the entrepreneurial educator. The unbundling of education opens the gate for entrepreneurs.

How could the DTA answer questions?

• FAQ and mentor DB as resources.
• A simple chatbot using the keyword method.
• The algorithm described here simply looks for keywords in the question.
• Then it responds with prepared answers connected to the keywords it finds.
• The chatbot can take any form: a human, an owl, a parrot.  
• I am inclined to use a dog.  
• The Digital Teaching Assistant (DTA) for MUVES (virtual worlds): Summary

*****************************

FAQ database for a bot

• The FAQ DB needs:
• 1. a FAQDB of frequently asked questions with answers.
• 2. a DB of keywords matched with the relevant answers
• 3. a DB of keyword synonyms matched with the keywords.
•  A  list of guiding comments the bot can add.

Building the FAQ DB

• Any SME can supply an initial set of FAQ and keywords.
• I recommend a brainstorming session to gather more extensive lists.
• The online thesaurus is good for synonyms.
• Guiding questions might mention categories of key words such as STEM subjects.

Maintaining the FAQ DB

• Questions the DTA can't answer go into the NAQ (Newly asked questions) DB.
• Editorial review will find additional FAQ in this set.
• Access records can identify infrequent items

Plan for interface function

1. Interface function (INT) takes typed questions on click and delivers them to bot
2. If bot returns an empty result INT sends the question to mentor search with NO KEY tag and apologizes.
3. Else: INT displays the result and asks if the user is satisfied.  
4. If the user is not satisfied, INT sends the question to Mentor search 
5. If the user is satisfied, the program ends.

Plan for bot

1. Bot parser checks the input for synonyms and replaces them with keywords.
2. Bot parser lists keywords in the input and removes duplicates.
3. Bot parser delivers Klist to Botwork
4. Botwork: If Klist empty, deliver input to mentor search. End function
5. Else: Use keywords to collect answers.  Deliver answers (as text string) and Klist to INT.
6. Add guidance if indicated (TBD)
7. End function.

Plan for Mentor search (Ms)

• If input has NO KEY tag, Ms sends the question to the wide spectrum mentors. 
• Else: Ms sends the question to mentors with KeyWords matching Klist.  

Plan for mentors

• The task of a mentor is to find out how to fix the system so that it can answer the question  the mentor received,

The flipped Webinar

• The flipped webinar of virtual worlds, a place for exchanging information. No cost, no reservation. Come as you are. Meet the online educator community.
• Flipped webinar for educators. A distributed conference. Short presentations, followed by focused discussions. With plans for outreach.
• The flipped webinar: Discovering online learning. The potential is huge, but you gotta know the territory.
• Planning for the flipped and persistent webinar. Guided by the participants. Informed by a panel. Organized by 3DWebWorldz

*********************************

Visit me on the web

• Drop by my web offices Weekdays: 12:-12:30 pm Central time (US)
• I am available for free consulting on any topic in this blog.
• Cybalounge and 3DWebWorldz (Orientation room)
• I will be in both places, so you may need to speak to get my attention.
• Web-worlds, 3D virtual worlds running in a browser. Summary
• And we can visit the Writer's Workshop on the Web
• Don't register -- enter as guest.  
• Due to medical issues, I cannot promise to be there.  But I will try.

************************

License

• Original text in this blog is CC By: unless specified public domain.
• Use as you please with attribution: link to the original.
• All images without attribution in this blog are CC0: public domain.
• Screenshots from ShareX
• Second Life, Linden, SLurl, and SL are trademarks of Linden Research Inc.
• Annotated screenshots made with Techsmith Capture
• This blog is not affiliated with anything.   Ads are from Google.
• Selby Evans in Kitely and Hypergrid, Thinkerer Melville in Second Life.