Quantum inspired qubit qutrit neural networks for real time financial forecasting

A new research paper by Kanishk Bakshi and Kathiravan Srinivasan, published in *Scientific Reports*, demonstrates a significant leap in AI for finance. The study compares three machine learning models for stock prediction: a classical Artificial Neural Network (ANN), a Quantum Qubit-based Neural Network (QQBN), and their novel Quantum Qutrit-based Neural Network (QQTN). While all models showed robust accuracy above 70%, the QQTN consistently emerged as the top performer, surpassing its counterparts in key quantitative metrics.

The QQTN's advantages are both in performance and efficiency. It demonstrated superior risk-adjusted returns, as measured by the Sharpe ratio, and greater prediction consistency via the Information Coefficient. Furthermore, it proved more robust under varying market conditions. Perhaps most compelling for practical deployment is its efficiency: the QQTN achieved this comparable or superior performance with "significantly reduced training times." This combination of high accuracy, consistency, and speed is precisely what's needed for real-time financial forecasting and high-frequency trading applications.

The research, categorized under both Artificial Intelligence and Quantum Physics, represents a 'quantum-inspired' approach. This means it uses principles from quantum computing—like leveraging qutrits (three-state quantum systems) instead of just qubits (two-state)—to create more powerful classical neural network architectures. The authors argue this work paves the way for integrating such advanced, computationally efficient models into other data-intensive fields beyond finance, highlighting the transformative potential of cross-disciplinary AI research.