fintech-kMC: Agent based simulations of financial platforms for design and testing of machine learning systems

Paper: arXiv 2301.01807

Authors: Isaac Tamblyn, Tengkai Yu, Ian Benlolo

Abstract

We discuss our simulation tool, fintech-kMC, which is designed to generate synthetic data for machine learning model development and testing. fintech-kMC is an agent-based model driven by a kinetic Monte Carlo (a.k.a. continuous time Monte Carlo) engine which simulates the behaviour of customers using an online digital financial platform. The tool provides an interpretable, reproducible, and realistic way of generating synthetic data which can be used to validate and test AI/ML models and pipelines to be used in real-world customer-facing financial applications.

Complexity vs Empirical Score

• Math Complexity: 5.5/10
• Empirical Rigor: 4.5/10
• Quadrant: Lab Rats — theoretically deep, empirically untested

Why this score: The paper introduces an innovative kinetic Monte Carlo agent-based framework for generating synthetic user data on digital financial platforms. While mathematically sound conceptually, it lacks deep empirical validation against real-world platform transaction data to fully prove its rigour.

Research Flowchart

  flowchart TD
    A["Research Goal:<br>Generate realistic synthetic data for financial ML testing"] --> B["Inputs & Parameters:<br>Platform rules, customer behaviors & agent profiles"]
    B --> C["Methodology:<br>Agent-Based Model + Kinetic Monte Carlo Engine"]
    C --> D["Computational Process:<br>Continuous-time simulation of customer platform usage"]
    D --> E["Outcomes:<br>Interpretable, reproducible & realistic synthetic data"]
    E --> F["Key Application:<br>Robust validation & testing of financial AI/ML pipelines"]