franciscovilchezv / chatbot-covid-generator

CSCE 771 - NLP Project

Hello, I'm Francisco 👋

A software developer and a person who loves to share knowledge.

Here is a list of materials that I have used in my classes or just when I had to explain something.

• Platform Based Development
• Introduction to Computer Science
• UNIX/Linux Fundamentals

I also have some guides regarding how to use different tools

• Using pass
• RabbitMQ terminology

You can find more updated information about myself in my website.

Thanks!

chatbot-covid-generator's Issues

Sept 3rd, 2020 - Meeting Notes

• (won't do) Read following tutorial by Paul Crook, FB on different methods for chatbot creation

Additional comments included by the professor while sharing the link:

• DialogFlow is the easiest to build chatbots. They allow slot-filling. Start with it.

• Also, try out MindMeld (open source) if interested.

• ParlAI is the one we can focus on. It is a RL-based dialog framework and can scale to small and big cases. 

https://
parl.ai/


• 
IBM Watson Conversation

• Read the following paper regarding the Motivation for chatbot generator.

Own notes:
Three ways of creating chatbots explained during the meeting by the professor:

• Finite automata (easiest)
• Slot filling (recommended). e.g. dialogflow
• Learning based. e.g parl.ai (open source)

At least cover 2 languages (e.g. English & Spanish)

• (splitted into separate issue) Build a chatbot

Website with covid information gathered by the professor

TODO:

• Fill proposal in the spreadsheet/doc from email

Creation of a generative based chatbot

Based on this tutorial:

https://heartbeat.fritz.ai/building-a-conversational-chatbot-with-nltk-and-tensorflow-part-1-f452ce1756e5

https://heartbeat.fritz.ai/building-a-conversational-chatbot-with-nltk-and-tensorflow-part-2-c67b67d8ebb

Creation of basic chatbot from scratch

Follows the following tutorial:

https://medium.com/analytics-vidhya/building-a-simple-chatbot-in-python-using-nltk-7c8c8215ac6e

Oct 23rd, 2020 - Meeting notes

Project was presented to Prof. Biplav with the functionalities as of commit 42c15fe. Following observations were made:

• Enhance prompt to differentiate User from Bot:
  User $:
System (Covid:ES:Mexico)$:

• Fix accuracy of chatbot for generic/specific questions.

• Turn everything into an infinite loop with options like Go back with 'Bye' to return to generation loop.

• Welcome message based on language

• Expand answers with "Show one answer. Say showing 1 of N." message and probably 'continue' option.

• Project presentation will be 1 slide (5 minutes). Demo should be shown

• Project report

Regarding paper presentation:

• 1 slide presentation (5 mins). 1 paper from:
  acl 2019, acl 2020, aaai 2019, aaai 2020, ijcai 2019, ijcai 2020

Generation of Chatbot for covid dataset in English