An Explainable Probabilistic Classifier for Categorical Data Inspired to Quantum Physics
The task of data classification in various contexts requires innovative machine learning methods. Categorical data are heterogeneous in terms of size, structural differences, and noise. That makes its representation in feature space non-trivial and time-consuming. Also, there is a growing demand for explainable and interpretable models. A recent paper suggests a classification algorithm for categorical data inspired by the superposition of states in quantum physics. ![]()
Image credit: TheDigitalArtist via Pixabay, free licence The researchers introduce the concept of wave-particle duality in machine learning. A generalized framework is proposed to unify the classical and the quantum probability. These new notions are used to create a new supervised classification algorithm. The suggested method achieves state-of-the-art performances without relying on data pre-processing and hyper-parameter tuning and provides a meaningful explanation of classification results. This paper presents Sparse Tensor Classifier (STC), a supervised classification algorithm for categorical data inspired by the notion of superposition of states in quantum physics. By regarding an observation as a superposition of features, we introduce the concept of wave-particle duality in machine learning and propose a generalized framework that unifies the classical and the quantum probability. We show that STC possesses a wide range of desirable properties not available in most other machine learning methods but it is at the same time exceptionally easy to comprehend and use. Empirical evaluation of STC on structured data and text classification demonstrates that our methodology achieves state-of-the-art performances compared to both standard classifiers and deep learning, at the additional benefit of requiring minimal data pre-processing and hyper-parameter tuning. Moreover, STC provides a native explanation of its predictions both for single instances and for each target label globally. Research paper: Guidotti, E. and Ferrara, A., "An Explainable Probabilistic Classifier for Categorical Data Inspired to Quantum Physics", 2021. Link: https://arxiv.org/abs/2105.13988
Image credit: TheDigitalArtist via Pixabay, free licence The researchers introduce the concept of wave-particle duality in machine learning. A generalized framework is proposed to unify the classical and the quantum probability. These new notions are used to create a new supervised classification algorithm. The suggested method achieves state-of-the-art performances without relying on data pre-processing and hyper-parameter tuning and provides a meaningful explanation of classification results. This paper presents Sparse Tensor Classifier (STC), a supervised classification algorithm for categorical data inspired by the notion of superposition of states in quantum physics. By regarding an observation as a superposition of features, we introduce the concept of wave-particle duality in machine learning and propose a generalized framework that unifies the classical and the quantum probability. We show that STC possesses a wide range of desirable properties not available in most other machine learning methods but it is at the same time exceptionally easy to comprehend and use. Empirical evaluation of STC on structured data and text classification demonstrates that our methodology achieves state-of-the-art performances compared to both standard classifiers and deep learning, at the additional benefit of requiring minimal data pre-processing and hyper-parameter tuning. Moreover, STC provides a native explanation of its predictions both for single instances and for each target label globally. Research paper: Guidotti, E. and Ferrara, A., "An Explainable Probabilistic Classifier for Categorical Data Inspired to Quantum Physics", 2021. Link: https://arxiv.org/abs/2105.13988