• Overview
  • Software Components
  • Target Audience
  • Prerequisites
  • Hardware Requirements
• Getting Started
  • Installation System Requirements
  • Obtain API key
  • Clone The Repository
  • Set up the environment file
    • API Key Environment variable
    • Conda Environment
    • Authenticate Docker with NGC
  • Running the workflow
  • Customizing the Workflow
• License
• Terms of Use

Notice: This README is for users running the notebook locally and makes assumptions that the software can be installed on the hardware.

Financial losses from worldwide credit card transaction fraud are projected to reach more than $403 billion over the next decade. Transaction fraud poses a major challenge for financial institutions, which struggle to detect and prevent increasingly complicated fraudulent activities. Traditional fraud detection methods, which rely on rules-based systems or statistical methods, are reactive and increasingly ineffective in identifying sophisticated fraudulent activities. As data volumes grow and fraud tactics evolve, financial institutions need more proactive, intelligent approaches to detect and prevent fraudulent transactions.

This NVIDIA AI Blueprint provides a reference example to detect and prevent sophisticated fraudulent activities for financial services with high accuracy and reduced false positives. It shows developers how to build a financial fraud detection workflow using the NVIDIA container for fraud detection. For model building, the Financial Fraud Training container augments fraud detection using graph neural networks (GNNs)—a deep learning technique—for improved accuracy. Inference is done using NVIDIA Dynamo-Triton (formerly Triton Inference Server) and produces fraud scores along with Shapley values for explainability. Furthermore, to help simplify the workflow, the Financial Fraud Training container also produces all the needed configuration files required by Dynamo-Triton.

This NVIDIA AI blueprint is broken down into three steps, which map to processes within a typical payment processing environment, those steps are: (1) Data Preparation, (2) Model Building, and (3) Data Inference. For this example, the data is a collection of files containing synthetic data. Within a production system, the event data is often saved within a database or a data lake. The data is prepared and then fed into the financial-fraud-training container model-building container. The output of the NIM folder with all the artifacts needs to be passed to Dynamo-Triton for inference.

This blueprint does not use any NVIDIA hosted services and runs fully in a locally hosted docker environment.

• financial-fraud-training container
• NVIDIA Dynamo-Triton

• Operating System: Ubuntu 20.04 or newer
• NVIDIA Driver version: 535 or newer
• NVIDIA CUDA version: 12.4 or newer
• NVIDIA Container Toolkit version: 1.15.0 or newer
• Docker version: Docker version 26 or newer

Everything is run via a Jupyter Notebook.

This Blueprint targets users that:

• understand the financial fraud space
• understand how to deploy container-based microservices
• understands how to run a Jupyter notebook

This notebook is a simple example of how to orchestrate a financial fraud detection workflow that leverages the financial-fraud-training container. The notebook uses a synthetic dataset and produces the accuracy and a confusion matrix. Using real data in a production environment would not alter the workflow.

• Obtain NVIDIA NVAIE key
• CUDA 12.5+ drivers installed

• GPU: 1x A6000, A100, or H100, minimum of 32 GB of memory
• CPU: x86_64 architecture
• Storage: 10 GB
• System Memory: 16 GB

• git
• Jupyter Notebook or Jupyter Lab

Additional required Python packages are installed in the conda environment step.

Here are two possible methods to generate an API key for NGC:

• Sign in to the NVIDIA Build portal with your email.
• Sign in to the NVIDIA NGC portal with your email.
  • Select your organization from the dropdown menu after logging in. You must select an organization which has NVIDIA AI Enterprise (NVAIE) enabled.
  • Click on your account in the top right, select "Setup" from the dropdown.
  • Click the "Generate Personal Key" option and then the "+ Generate Personal Key" button to create your API key.
    • This will be used in the NVIDIA_API_KEY environment variable.
    • Click the "Generate API Key" option and then the "+ Generate API Key" button to create the API key.

IMPORTANT: This will be used in the NVIDIA_API_KEY environment variable below.

• API catalog keys: NVIDIA API catalog or NGC

git clone https://github.com/NVIDIA-AI-Blueprints/Financial-Fraud-Detection

# Create the API environment
export NVIDIA_API_KEY=your_key

Note: Add the environment variable to the .bashrc file to persist it beyond a single instance.

# to persist the key
echo "export NVIDIA_API_KEY=your_key" >> ~./bashrc

The workflow uses Conda to create an environment with all the needed packages. You can get a minimum installation of Conda and Mamba using Miniforge.

Create an environment using the following command, making sure that you are in the Financial-Fraud-Detection folder.

 mamba env create -f conda/notebook_env.yaml

Finally, activate the environment.

conda activate notebook_env

In order to pull images required by the Blueprint from NGC, you must first authenticate Docker with NGC. You can use same the NVIDIA API Key obtained in the Obtain API key section (saved in the NVIDIA_API_KEY environment variable).

echo "${NVIDIA_API_KEY}" | docker login nvcr.io -u '$oauthtoken' --password-stdin

No command line option (CLI) is available without converting a Jupyter Notebook to a Python file.

Prerequisite: Start a browser

Starting Jupyter

Option 1:

   jupyter notebook

Option 2:

Starting Jupyter Notebook with the notebook:

   cd notebook
   jupyter notebook financial-fraud-usage.ipynb

Option 3:

Starting Jupyter Labs

jupyter-lab

or 

jupyter-lab --ip=* --no-browser

In either case above, Jupyter will output status information. One key line is the URL:

$ jupyter notebook
   ...
   The Jupyter Notebook is running at: http://localhost:8888/

Via the browser, connect to the specified URL and process the notebook.

The NIM offers a number of customization options. Detail on those can be found here (*** NEED LINK).

By using this software or microservice, you are agreeing to the terms and conditions of the license and acceptable use policy.

GOVERNING TERMS: The NIM container is governed by the NVIDIA Software License Agreement and Product-Specific Terms for AI Products.