[Day 200] Kukmin Bank AI competition project idea

 Hello :)
Today is Day 200!


A quick summary of today:
  • creating a project plan, timeline and learning about graph DBs


My lab mate and I considered two project ideas. First I will talk and show our plan for the project we chose, and then for the other one which I did a simple demo for.

Project idea we chose: Real-time fraud analysis using graph data and graph database

My lab mate said he trusts me to choose the technologies for our project and to set up a plan to follow:

Training Pipeline

Data Collection and Preprocessing:

  • Raw dataset obtained from Kaggle containing transaction data.
  • Raw data is preprocessed before going into the database.

Storage in ArangoDB:

  • Transactions are stored in ArangoDB, a popular graph database.
  • Customers and merchants are represented as nodes, and transactions as edges.

Graph Neural Network with PyG:

  • PyTorch Geometric (PyG) is used to build and train a Graph Neural Network (GNN) for fraud detection on the graph data.
  • The model is trained on a batch of the dataset and saved for inference.

Real-time inference pipeline

Simulated Data Streaming:
  • Emulating a live transaction environment with continuous data flow using the downloaded full Kaggle dataset.
Data Ingestion with Kafka:
  • Kafka handles the real-time streaming of transaction data, ensuring efficient data flow and availability.
Storage in ArangoDB:
  • Real-time transactions are stored in ArangoDB, maintaining the graph structure.
Real-Time Fraud Detection:
  • The trained GNN model analyses new transactions in real-time as they are streamed and stored.
  • The model uses the saved GNN model to perform real-time inference and detect fraudulent transactions.

Also in excel, I used a project timeline template to create:
The competition started on Wednesday and submission deadline is 11th of Aug, so we are starting from today. The first task is finding a decent dataset with a decent amount of transactions.

One of the new things for me is using a graph db. I saw there is neo4j nd this ArangoDB, I decided to use ArangoDB because the python setup was easier. From playing with it, I included sample transactions and the db webapp (similar to postgres' pgAdmin) and got a limited (because there are too many nodes and edges) visualisation:

I am really looking forward to this project and working with my lab mate - Jae-hyeok Choi.  


As for the idea that we dismissed - banking voice assistant

I was up for both, and my lab mate preferred the first one, and now looking back - I am glad haha because it is so exciting.

Nevertheless, to set up a voice assistant demo was surprisingly easy. I set it up using huggingface spaces. Here is the link.

The above ~30 lines of code take voice -> turn it into text -> a language model answers the text -> the answer is transformed to speech and returned to the user. To try the demo, one needs an OpenAI api key. In order for this simple voice assistant to a bank once, I need to use the LM for a RAG app that talks to a db. Given I made db2chat it should not be that difficult. But this is a project for later. 


That is all for today!

See you tomorrow :) 

Popular posts from this blog

[Day 198] Transactions Data Streaming Pipeline Porject [v1 completed]

Image
 Hello :) Today is Day 198! A quick summary of today: data streaming pipeline project [v1 done] Here is a link to the project's repo. Well ... I did not know I can do it in a day (~14 hours) after yesterday's issues but here we are. Turns out in order to insert the full (~70 variables with nested/list structure), I need the proper pyspark schema. And yesterday I did not have that and that is why when I was reading data in the kafka producer I was getting NULL in the columns - my schema was wrong. Well today I not only fixed the schema for the 4 variables I had yesterday, but included *all* the variables that come from the Stripe API ~ 70 (for completeness).  When I run docker-compose, the data streams and is input into the postgres db (and is still running). Unfortunately, the free Stripe API for creating realistic transactions has a limit of 25, so every 3 seconds, 25 new transactions are sent to the db. It has been running half the day (since I got that set up) and as I am wr
Read more

[Day 107] Transforming natural language to charts

Image
 Hello :) Today is Day 107! A quick summary of today: created text2chart - transforming natural language to charts [ github ] [ webapp ] I am super sick, but in the around 3 hours of my fever not being high, I managed to study a bit. I was again thinking about creating some nlp project but was stuck on the idea of how to host a model. After looking around I found this article where they create a similar app called chat2vis  and they use codellama. I know codellama is an open source model that I can get from huggingface and I wondered, how did they host it. Turns out - they ask the user to input a huggingface api key, which then allows for querying and generating content. (AWESOME! this is applicable to some of the other PDF chat apps that I have, but I will look to make an app of them later).  My fever eased a bit, and I sat on my chair to do some work ^^ Here is a breakdown of the app. I used streamlit. Also, 2 default datasets during dev: Financial Statements of Major Companies(2009-
Read more

[Day 54] I became a backprop ninja! (woohoo)

Image
 Hello :) Today is Day 54! A quick summary of today: Read through cs231n's neural network lectures [1] [2] [3] Spent 8 hours going over Andrej Karpathy's backprop ninja tutorial After yesterday's confidence boost in doing backprop, I came in today with the intention of becoming of, as Andrej Karpathy calls it, a backprop ninja.  The cs231n NN lectures were the usual about how NN became to be, param initialization, activation functions, loss functions, regularization and optimizers. Nothing out of the ordinary, but still was good to reassure myself of that knowledge. But that was definitely not the main even of today. As I mentioned, I wanted to become (or at least attempt to become) a backprop ninja today. I have watched that particular tutorial twice, the 1st time I had no idea what is going on, the 2nd time I could follow, and now it was the 3rd time - and now, I wanted to participate and actually see the why and the how of the things we have to do, when we do manual back
Read more