Inspiration
ForwardSentry was inspired by the growing need to understand and track the spread of information/mis-information in our connected world. We recognized that the same principles used to trace message forwarding could be applied to track financial transactions, potentially uncovering complex networks of communication or money laundering schemes.
What it does
ForwardSentry is a sophisticated application that traces the path of forwarded messages across networks. It uses Arango graph database technology and semantic analysis to:
- Map the journey of messages as they're forwarded between users
- Identify similar messages using vector embedding similarity
- Using web, verifies if the sent message was a fact or fake.
How we built it
Built ForwardSentry using:
- ArangoDB as our graph database to store and query relationships between users and messages
- Python for backend logic and data processing
- Sentence transformers for semantic analysis of message content
- SerpAPI for web search to verify authenticity of message.
- AQL (ArangoDB Query Language) for efficient graph traversal and similarity searches
Challenges we ran into
- Designing an efficient graph schema that could represent both message forwarding and financial transactions
- Implementing semantic search functionality that was both accurate and performant
Accomplishments that we're proud of
I am glad to convert my vision into working proof-of-concept and i am excited to explore more and build further.
What we learned
Learned about ArangoDB functionalities and its features, how to build a RAG app with graph database
What's next for ForwardSentry
Introducing further features and fine-tuning existing ones.
Built With
- arangodb
- langchain
- python
- serpapi
Log in or sign up for Devpost to join the conversation.