Webinar recording
In this webinar session, we show how GraphAware Hume can leverage GPT and LLM technology to accelerate analysis of unstructured datasets, with a specific focus on law enforcement.
Our live demonstration centres on public judicial reports detailing a vast web of state corruption in South Africa, commonly known as the “state capture” or the Gupta Leaks.
By employing OpenAI’s GPT (Generative Pre-trained Transformers), we will showcase how the model facilitates named entity recognition (NER) and relation extraction (RE) to extract pertinent knowledge from these texts.
You will learn
- Prompt engineering tips
Use GPT to identify key elements that are crucial for knowledge graph creation, including entities and the relationships between them. - LLMs in action
Explore the use of large language models (LLMs) for building knowledge graphs, including data cleaning and normalisation processes - Results analysis
Investigate the central questions surrounding the Gupta Leaks:- How did the Guptas sustain corruption on this scale for an extended period without arousing suspicion?
- Which organisations are linked to the Guptas?