Generative AI Hackathon with IBM Granite

Start

Emerald

CS Edwards

Utilized IBM granite instruct and code models to extract themes and relationships from documents.

Intuitive visual representation of document data
Easily identify key themes and connections for  research and analysis
Downstream tasks:
Recommendation Systems
Data Integration
Semantic Search
Industries:
Education
Finance
Policy
More…

Successfully extracted themes and relationships using Granite model and generated Cypher queries for knowledge graph generation in Neo4j


Challenges:
Given the non-deterministic nature of LLMs Granite code instruct does not always generate a valid query.
Future Work:
Refining prompts for better/ more predictable and structured outputs from model


Utilized IBM granite instruct and code models to extract themes and relationships from documents

"Emerald team on Generative AI Hackathon with IBM Granite Hackathon"

Team Idea

Submission