Gated Mixture Variational Autoencoders for Value Added Tax audit case selection

  title={Gated Mixture Variational Autoencoders for Value Added Tax audit case selection},
  author={Christos Kleanthous and Sotirios P. Chatzis},
  journal={Knowl. Based Syst.},
A Decision Support System for Corporate Tax Arrears Prediction
The best performing decision support system was a hybrid based on the random forest method for observations with previous tax arrears in at least two months and a logical rule for the rest of the observations.
Association Rules and Machine Learning for Enhancing Undeclared Work Detection
  • Eleni Alogogianni, M. Virvou
  • Business
    2020 11th International Conference on Information, Intelligence, Systems and Applications (IISA
  • 2020
An advanced data analysis method, the association rule mining, is demonstrated, which has significant advantages over rule-based systems, in classifying employers likely to engage in undeclared work and revealing insights in patterns of employers’ illegal behaviour that were previously unidentified.
Data Science Methods and Techniques for Goods and Services Trading Taxation: a Systematic Mapping Study
A systematic mapping of the literature aimed to identify how data science methods and techniques have been applied to this context and how the problems inherent in this domain are being handled and results show that data science can efficiently be used to improve the collection of these types of taxes.
O Uso de Inteligência Artificial no Combate à Evasão Fiscal: Uma Revisão Sistemática da Literatura
Tax evasion is a problem faced by governments around the world. The use of AI has been a viable alternative to combat this problem. In this context, this work aims to identify how AI helps to combat
A general Neural Particle Method for hydrodynamics modeling


Asymmetric deep generative models
Tax Fraud Detection for Under-Reporting Declarations Using an Unsupervised Machine Learning Approach
The ability of the model to identify under-reporting taxpayers on real tax payment declarations is demonstrated, reducing the number of potential fraudulent tax payers to audit and increasing the operational efficiency in the tax supervision process without needing historic labeled data.
Adaptive Subgradient Methods for Online Learning and Stochastic Optimization
This work describes and analyze an apparatus for adaptively modifying the proximal function, which significantly simplifies setting a learning rate and results in regret guarantees that are provably as good as the best proximal functions that can be chosen in hindsight.
Dropout: a simple way to prevent neural networks from overfitting
It is shown that dropout improves the performance of neural networks on supervised learning tasks in vision, speech recognition, document classification and computational biology, obtaining state-of-the-art results on many benchmark data sets.
Signal Modeling and Classification Using a Robust Latent Space Model Based on $t$ Distributions
A Bayesian approach to factor analysis modeling based on Student's-t distributions is developed, which provides an efficient and more robust alternative to EM-based methods, resolving their singularity and overfitting proneness problems, while allowing for the automatic determination of the optimal model size.
Deep Learning
Deep learning is making major advances in solving problems that have resisted the best attempts of the artificial intelligence community for many years, and will have many more successes in the near future because it requires very little engineering by hand and can easily take advantage of increases in the amount of available computation and data.