Skip to content

usercontext/DynamicEnrichment

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 

Repository files navigation

Dynamic Enrichment

Introduction

The are several knowledge graphs in form of hierarchies of categories present on the world wide web. Most of them belong to information sharing websites. These include product category hierarchies like those of Amazon, or task hierarchies like that of Wikihow. These hierarchies are very rich in information and contains several insights to how certain categories should be arranged. Internet is ever growing and categories of products, tasks etc. are increasing rapidly. However, these additions over time are accounted for, manually in these knowledge bases. It is a very tedious process to handle these manual changes due to the massive information they hold and the strict guidelines one has to follow.

We propose a novel concept which aims towards enriching such knowledge bases based on the prior information they contain. Previous works include automatic hierarchy generation based on the semantics. However, we aim at inferring knowledge from the existing hierarchy by observing the clustering strategy used by human experts and extrapolating that strategy to new incoming categories. This way we ensure automatic enrichment of hierarchy with the information flow.

Methods

Bayesian Rose Trees Implementation

Unsupervised Decision Trees

Hierarchical Reinforcement Learning

Learning Clustering Strategy using Reinforcement Learning

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published