Enhancing Community Interactions with Data-Driven Chatbots--The DBpedia Chatbot
0 likes129 views
The document describes the DBpedia Chatbot, a knowledge-graph driven chatbot that addresses four main challenges: understanding user queries, fetching relevant information from queries, tailoring responses to different platforms, and developing subsequent user interactions. It has an architecture that includes intent classification, relevance scoring to rank properties in knowledge cards for a given DBpedia class. The DBpedia Chatbot application and GitHub repo are provided.
![Enhancing Community Interactions with Data-Driven Chatbots - The DBpedia Chatbot
by Ram G Athreya, Axel-Cyrille Ngonga Ngomo and Ricardo Usbeck
Introduction
The DBpedia Chatbot is a knowledge-graph-
driven chatbot.
Four main challenges were addressed when
building the chatbot:
(1) Understanding user queries
(2) Fetching relevant information based on
the queries
(3) Tailoring the responses based on the
standards of each output platform (i.e. Web,
Slack, Facebook)
(4) Developing subsequent user interactions
with the DBpedia Chatbot
Architecture
Query Fulfillment Workflow
Screenshot (across platforms)
Intent Classification Relevance Score
Ranking properties in Knowledge Cards
For a given DBpedia Class
(1) We found the total number of occurrences of
that class in the knowledge graph.
(2) We extracted all rdfs:domain properties for
the respective class
(3) Calculated the number of distinct occurrences
of each individual property in the knowledge
graph
(4) We derived a relevance score r ∈ [0, 1]
for each property p for given class c:
Application:
http://chat.dbpedia.org
GitHub Repo:
https://github.com/dbpedia/chatbot](https://image.slidesharecdn.com/posterfinal-180718144902/85/Enhancing-Community-Interactions-with-Data-Driven-Chatbots-The-DBpedia-Chatbot-1-320.jpg)
Enhancing Community Interactions with Data-Driven Chatbots--The DBpedia Chatbot
- 1. Enhancing Community Interactions with Data-Driven Chatbots - The DBpedia Chatbot by Ram G Athreya, Axel-Cyrille Ngonga Ngomo and Ricardo Usbeck Introduction The DBpedia Chatbot is a knowledge-graph- driven chatbot. Four main challenges were addressed when building the chatbot: (1) Understanding user queries (2) Fetching relevant information based on the queries (3) Tailoring the responses based on the standards of each output platform (i.e. Web, Slack, Facebook) (4) Developing subsequent user interactions with the DBpedia Chatbot Architecture Query Fulfillment Workflow Screenshot (across platforms) Intent Classification Relevance Score Ranking properties in Knowledge Cards For a given DBpedia Class (1) We found the total number of occurrences of that class in the knowledge graph. (2) We extracted all rdfs:domain properties for the respective class (3) Calculated the number of distinct occurrences of each individual property in the knowledge graph (4) We derived a relevance score r ∈ [0, 1] for each property p for given class c: Application: http://chat.dbpedia.org GitHub Repo: https://github.com/dbpedia/chatbot