Fabric of Data
- Categories: #data #embeddings #integration #discovery #unstructured
Overview
Large volumes of data are represented in unstructured formats such as text. Often times, we want to access the knowledge of unstructured data in a structured format so we can ask questions more easily. In the fabric of data project we are exploring techniques, algorithms, and systems to manipulate unstructured and structured data more easily. We are building a relational embedding that helps us ask hard discovery queries over structured data. We are building text-table systems that help us ask questions in natural language over unstructured and structured sources as well as extract structure from text.