Building a Knowledge Base Question Answering Pipeline

Note: Sign-ups to this event are managed through Skills Matter. Please signup through their event page: https://skillsmatter.com/meetups/12073-building-a-knowledge-base-question-answering-pipeline

Knowledge base question answering aims to provide a natural language interface to factual knowledge. It requires precise modeling of the question meaning through the entities and relations available in the knowledge base in order to retrieve the correct answer.

In this session, we will discuss a semantic parsing approach to knowledge base question answering and the challenges of building a question answering pipeline. It is common to break down the task into three main steps: entity linking, relation disambiguation and answer retrieval. We will focus on two aspects: entity linking across various categories of entities and using graph neural networks to jointly encode entities and relations.

Papers:

• Entity Linking http://aclweb.org/anthology/S18-2007
• Graph Neural Networks for Semantic Parsing http://aclweb.org/anthology/C18-1280

Code:

• https://github.com/UKPLab/starsem2018-entity-linking
• https://github.com/UKPLab/coling2018-graph-neural-networks-question-answering

Further Reading:

• Semantic Parsing by Staged Query Generation http://aclweb.org/anthology/P15-1128

--

A note about the Journal Club format:

1. There is no speaker at Journal Club.
2. There is NO speaker at Journal Club.
3. We split into small groups of 6 people and discuss the papers. For the first hour the groups are random to make sure everyone is on the same page. Afterwards we split into blog/paper/code groups to go deeper.
4. Volunteers sometimes seed the discussion by guiding through the paper highlights for 5 mins. You are very welcome to volunteer in the comments.
5. Reading the materials in advance is really helpful. If you don't have time, please come anyway. We need this group to learn together.