Continual Learning for Unsupervised Anomaly Detection in Continuous Auditing of Financial Accounting Data (2021-12-25T00:00:00.000000Z)

TL;DR

This work proposes a continual anomaly detection framework to overcome both challenges and designed to learn from a stream of journal entry data experiences, and provides initial evidence that such a learning scheme offers the ability to reduce false-positive alerts and false-negative decisions.

Abstract

International audit standards require the direct assessment of a financial statement's underlying accounting journal entries. Driven by advances in artificial intelligence, deep-learning inspired audit techniques emerged to examine vast quantities of journal entry data. However, in regular audits, most of the proposed methods are applied to learn from a comparably stationary journal entry population, e.g., of a financial quarter or year. Ignoring situations where audit relevant distribution changes are not evident in the training data or become incrementally available over time. In contrast, in continuous auditing, deep-learning models are continually trained on a stream of recorded journal entries, e.g., of the last hour. Resulting in situations where previous knowledge interferes with new information and will be entirely overwritten. This work proposes a continual anomaly detection framework to overcome both challenges and designed to learn from a stream of journal entry data experiences. The framework is evaluated based on deliberately designed audit scenarios and two real-world datasets. Our experimental results provide initial evidence that such a learning scheme offers the ability to reduce false-positive alerts and false-negative decisions.

Authors

Marco Schreyer

3 papers

Damian Borth

3 papers

Hamed Hemati

1 papers

TL;DR

Abstract

Authors

References71 items

Multi-view contrastive self-supervised learning of accounting data representations for downstream audit tasks

Understanding Continual Learning Settings with Data Distribution Drift Analysis

Avalanche: an End-to-End Library for Continual Learning

Unsupervised anomaly detection for internal auditing: Literature review and research agenda

A Machine Learning Approach of Measuring Audit Quality: Evidence From China

Learning sampling in financial statement audits using vector quantised variational autoencoder neural networks

Lifelong Learning of Compositional Structures

Clinical applications of continual learning machine learning.

Learning from Machine Learning in Accounting and Assurance

A Continual Learning Survey: Defying Forgetting in Classification Tasks

Consideration of fraud in a financial statement audit

Applying Deep Learning to Audit Procedures: An Illustrative Framework

Lifelong Learning for Scene Recognition in Remote Sensing Images

Continual Learning with Tiny Episodic Memories

Reducing false positives in fraud detection: Combining the red flag approach with process mining

Experience Replay for Continual Learning

Don't forget, there is more than forgetting: new metrics for Continual Learning

Efficient Lifelong Learning with A-GEM

An Empirical Study of Example Forgetting during Deep Neural Network Learning

Selective Experience Replay for Lifelong Learning

Continual Lifelong Learning with Neural Networks: A Review

Detection of Anomalies in Large Scale Accounting Data using Deep Autoencoder Networks

Gradient Episodic Memory for Continual Learning

Continual Learning with Deep Generative Replay

Continual Learning Through Synaptic Intelligence

Manual journal entry testing: Data analytics and the risk of fraud

Overcoming catastrophic forgetting in neural networks

iCaRL: Incremental Classifier and Representation Learning

Lifelong Machine Learning

Learning without Forgetting

Progressive Neural Networks

Process Model Representation Layers for Financial Audits

Multilevel Process Mining for Financial Audits

Empirical Evaluation of Rectified Activations in Convolutional Network

Adam: A Method for Stochastic Optimization

Detecting Fraud Using Transaction Frequency Data

A Field Study on the Use of Process Mining of Event Logs as an Analytical Procedure in Auditing

A Review of ERP Research: A Future Agenda for Accounting Information Systems

A business process mining application for internal transaction fraud mitigation

Fraud Detection in ERP Systems Using Scenario Matching

Data mining journal entries for fraud detection: An exploratory study

Transaction mining for fraud detection in ERP Systems

Understanding the difficulty of training deep feedforward neural networks

Internal Fraud Risk Reduction - Results of a Data Mining Case Study

A Role Mining Inspired Approach to Representing U ser Behaviour in ERP Systems

SNARE: a link analytic system for graph labeling and risk detection

Computer Aided Fraud Prevention and Detection: A Step by Step Guide

Large Scale Detection of Irregularities in Accounting Data

Corporate Fraud Handbook: Prevention and Detection

Outlier Detection Using Replicator Neural Networks

Statistical fraud detection: A review

LOF: identifying density-based local outliers

Catastrophic forgetting in connectionist networks

A Case Study on the Application of Process Mining in Combination with Journal Entry Tests for Financial Auditing

Using Autoencoders for Data-Driven Analysis in Internal Auditing

Autoencoder Neural Networks versus External Auditors: Detecting Unusual Journal Entries in Financial Statement Audits

Journal entry anomaly detection model

Fighting Fraud: A never-ending Battle

Report to the Nations - Global Study on Occupational Fraud and Abuse

Data Mining Journal Entries for Fraud Detection: A Replication of Debreceny and Gray’s (2010) Techniques

Deep Learning

Auditing Journal Entries Using Extreme Value Theory

Auditing Journal Entries Using Self-Organizing Map

Cluster Analysis for Anomaly Detection in Accounting Data: An Audit Approach 1

International Standards on Auditing 240, The Auditor’s Responsibilities Relating to Fraud in an Audit of Financial Statements

Practice Aid for Testing Journal Entries and Other Adjustments Pursuant to AU Section 316

Supporting Online Material for Reducing the Dimensionality of Data with Neural Networks

Principles of Analytic Monitoring for Continuous Assurance

THE CONTINUOUS AUDIT OF ONLINE SYSTEMS

Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem

The Law of Anomalous Numbers

Field of Study

Journal Information

Name

Volume

Venue Information