NEO4J Generated Graph
WebApp Interface
Inspired by KETOPRAK (Indonesian traditional food)
User Diagram
Model Evaluation

KETOPRAK - Knowledge Extraction and Threat Operations Platform for Real-time Analysis and Knowledge

Inspiration

The inspiration behind KETOPRAK stemmed from the need to enhance internal security by leveraging advanced technologies to analyze and extract knowledge from unstructured text, particularly related to terrorism. The goal was to create a platform that converts reports and articles into a structured knowledge graph, enabling efficient querying and analysis through a chatbot interface.

What it does

KETOPRAK is a platform designed to convert unstructured reports and articles into a structured knowledge graph. It enables users to query and analyze data efficiently, providing real-time insights and responses through a chatbot interface. The system enhances internal security by facilitating the extraction and operationalization of knowledge related to security threats.

How we built it

We built KETOPRAK using a combination of modern technologies:

Frontend: Next.js
Backend: FastAPI
LLM Integration: LangChain
Database: Supabase for storing prompts and responses, and Neo4j for querying the knowledge graph

The architecture involves user interactions with the frontend, authentication and backend processing, querying the LLM, and storing data in the database.

Challenges we ran into

During the development of KETOPRAK, we faced several challenges:

Integrating various technologies into a cohesive system
Ensuring the accuracy and efficiency of the LLM in processing and extracting relevant data
Handling large volumes of unstructured text and converting it into a structured knowledge graph
Managing real-time data storage and retrieval for prompt and response handling

Accomplishments that we're proud of

We are proud of the following accomplishments:

Successfully integrating multiple technologies to create a functional platform
Developing an efficient and accurate LLM for knowledge extraction
Creating a seamless user interface for querying and analyzing data
Implementing a robust backend system for data processing and storage

What we learned

Throughout the development process, we learned:

The importance of seamless integration between frontend and backend technologies
Techniques for efficient knowledge extraction and operationalization using LLMs
Best practices for managing and querying large datasets in real-time
Strategies for ensuring system scalability and reliability

What's next for KETOPRAK

The future plans for KETOPRAK include:

Enhancing the LLM capabilities for more accurate and diverse knowledge extraction
Expanding the system to handle a wider range of security-related topics
Improving the user interface for better user experience
Implementing advanced security measures to protect sensitive data
Scaling the platform to support more users and handle larger datasets