Using graphs as basic representation of data for ML purposes has several advantages: (i) the data is already modeled for further analysis, explicitly representing connections and relationships between things and concepts; (ii) graphs can easily combine multiple sources into a single graph representation and learn over them, creating Knowledge Graphs; (iii) improving computation performances and quality. The talk will discuss these advantages and present applications in the context of recommendation engines and natural language processing.

In this talk, Christophe will describe a graph-centric cognitive computing pipeline and detail the process from the ingestion of unstructured text up to the generation of a knowledge graph, queryable using natural language through chatbots built with IBM Watson Conversation.