Quantum Machine Learning Enhances Financial Forecasting, Study by Itaú Unibanco Reveals

Quantum machine learning, a rapidly evolving field, has the potential to revolutionize financial forecasting, according to a study by researchers Sohum Thakkar, Skander Kazdaghli, Natansh Mathur, Iordanis Kerenidis, André J Ferreira Martins, and Samurai Brito. The study, conducted at Itaú Unibanco, Latin America’s largest bank, focused on improving Random Forest methods for churn prediction and neural network models for credit risk assessment using quantum algorithms. The results showed that quantum machine learning can enhance the performance of these models, with potential for further improvements as quantum hardware advances. Future research could explore more sophisticated quantum algorithms and their application in other financial areas.

What is Quantum Machine Learning and How Can it Improve Financial Forecasting?

Quantum machine learning is a rapidly evolving field that has the potential to revolutionize various domains, including finance. This technology uses quantum algorithms to enhance machine learning across a variety of applications. In a recent study, researchers demonstrated how quantum machine learning can be used to improve financial forecasting. The research was conducted by Sohum Thakkar, Skander Kazdaghli, Natansh Mathur, Iordanis Kerenidis, André J Ferreira Martins, and Samurai Brito, and was accepted for publication in March 2024.

The researchers focused on two use cases within the business of Itaú Unibanco, the largest bank in Latin America. The first use case aimed to improve the performance of Random Forest methods for churn prediction. The second use case explored the performance of neural network models for credit risk assessment. The results of the study demonstrated that leveraging quantum ideas can effectively enhance the performance of machine learning, both today as quantum-inspired classical ML solutions and even more in the future with the advent of better quantum hardware.

How Can Quantum Machine Learning Improve Churn Prediction?

In the first use case, the researchers aimed to improve the performance of Random Forest methods for churn prediction. They introduced quantum algorithms for Determinantal Point Processes (DPP) sampling and developed a method of DPP sampling to enhance Random Forest models. The researchers evaluated their model on the churn dataset using classical DPP sampling algorithms and performed experiments on a scaled-down version of the dataset using quantum algorithms.

The results demonstrated that in the classical setting, the proposed algorithms outperformed the baseline Random Forest in precision efficiency and bottom line. They also offered a precise understanding of how quantum computing can impact this kind of problem in the future. The quantum algorithm run on an IBM quantum processor gave similar results as the classical DPP on small batch dimensions but faltered as the dimensions grew bigger due to hardware noise.

How Can Quantum Machine Learning Improve Credit Risk Assessment?

In the second use case, the researchers aimed to explore the performance of neural network models for credit risk assessment by incorporating ideas from quantum compound neural networks. They started by using quantum orthogonal neural networks, which add the property of orthogonality for the trained model weights to avoid redundancy in the learned features. These orthogonal layers, which can be trained efficiently on a classical computer, are the simplest case of what the researchers call compound neural networks, which explore an exponential space in a structured way.

For their use case, the researchers designed compound neural network architectures that are appropriate for financial data. They evaluated their performance on a real-world dataset and showed that the quantum compound neural network models both have far fewer parameters and achieve better accuracy and generalization than classical fully connected neural networks.

What is the Future of Quantum Machine Learning in Finance?

The results of this study demonstrate the potential of quantum machine learning methods in improving the performance of forecasting in finance. However, the field is still rapidly evolving, and there are many potential future research directions. For example, the researchers suggest that future work could focus on developing more sophisticated quantum algorithms for DPP sampling, which could further improve the performance of Random Forest models for churn prediction.

Additionally, the researchers suggest that future work could explore the use of quantum compound neural networks for other financial applications. For example, these networks could potentially be used to improve the performance of machine learning models for predicting stock prices or for detecting fraudulent transactions.

In conclusion, while quantum machine learning is still a relatively new field, it holds great promise for improving the performance of financial forecasting and other applications in finance. As quantum hardware continues to improve, we can expect to see even more significant enhancements in the performance of machine learning models that leverage quantum ideas.