BlogQuantum Computing in Finance: A Leap Forward with Itaú UnibancoQC Ware7/20/2023

--ListenShareQuantum computing’s rapid evolution promises transformative potential for various sectors, notably within the realms of finance and machine learning. In a recently published paper in collaboration with Itaú Unibanco, Latin America’s largest bank, our team explored the impact of quantum machine learning on financial forecasting, spotlighting an exciting frontier for the financial industry.The emergence of quantum computing introduces substantial capabilities to tackle computational challenges in finance, from combinatorial optimization to Monte Carlo simulations, and advanced machine learning techniques. Itaú Unibanco, always at the forefront of technological advancements, partnered with us to dive deep into this promising field, focusing on two key applications: churn prediction and credit risk assessment.Addressing customer churn is a persistent challenge for numerous financial institutions. To tackle this, we’re utilizing the capabilities of quantum computing, both by providing pure quantum solutions that can be tested on the current quantum hardware and quantum-inspired solutions that can be put in production today. Our hardware experiments were performed on a 16-qubit IBM machine named ibm_guadalupe, while we used GPUs for the quantum-inspired full-scale demonstration.In our quest to bolster the effectiveness of Random Forest models, we have implemented a Determinantal Point Processes (DPP) sampling method, which was first developed for studying fermionic systems. This proposed DPP Random Forest model has shown remarkable improvements, increasing precision by a significant 6%.In collaboration with Itaú Unibanco, our approach has successfully outperformed traditional Random Forest techniques in both precision and efficiency. This achievement underscores quantum computing’s instrumental role in redefining and improving the future of churn prediction strategies.In the vital arena of credit risk assessment, our research was greatly enriched by Itaú Unibanco’s operational experience. Our approach incorporated Quantum Neural Networks, specifically using orthogonal and compound layers to eliminate redundancy in learned features. For our hardware experiments, we utilized the 27-qubit IBM machine ibm_hanoi, while we simulated the neural networks on the full data using GPUs.The efficiency of our quantum neural network is worth noting as it used an impressive 87.5% fewer parameters compared to classical models, without sacrificing accuracy. This leaner structure aids in improving model interpretability, which is a crucial factor in credit risk assessment models. Consequently, these quantum-enhanced models achieved notable advancements in both accuracy and generalization, offering a promising approach for the future of credit risk analysis.Our collaboration with Itaú Unibanco demonstrated that quantum machine learning presents a robust avenue for enhancing machine learning, both in the present and the future. As quantum hardware continues to evolve, the potential for even more significant breakthroughs grows.Together with Itaú Unibanco, we’ve illuminated a path into the future, where the formidable power of quantum computing can revolutionize key finance applications.For those interested in exploring the intricate details of our research, we encourage you to dive into the full paper.Are you ready to explore how quantum computing can benefit your organization? Start your quantum journey today. Get in touch with us and let’s create the future of finance together.