• We utilize Transformers and Quantum LSTM models to forecast time series of deposit rates. This decreases MAPE, one of the best error metrics, and enhances backtesting.

​​

            Credit Risk

​

• We increase the ROC in credit scoring using Quantum Convolutional Neural Network, improving the discriminant power and Backtesting.

​​

• We use Bayesian Neural Network to reduce the uncertainty in the forecasting of PD.

​​

• Use Deep Learning Survival and Random Forest Survival instead of Cox Regression to estimate lifetime PD improvement backtesting.

​​

• With noise, uncertainty, and lack of data, we utilize Robust Machine Learning to model LGD, reducing Model Risk.

​​

• The economic capital for credit risk has been estimated using Quantum Monte Carlo faster than Simulation Monte Carlo.

​​

            Counterparty Credit Risk

​

• We utilized a Quantum Neural Network to simulate paths for calculating the Credit Value Adjustment of a derivatives portfolio. The trained neural networks replace the original pricing model. 

            CyberRisk

​

• We explain how Shor's algorithm, which can factorize quickly on a quantum computer, undermines RSA's cryptography security assumptions. We also expose how Lattice-based constructions support standards of post-quantum cryptography. 

     

             Model Risk

​

• Expose the state-of-the-art methods in interpretable machine learning and model diagnosis.

​​

• Reduce the uncertainty in lifetime PD estimation using Quantum Markov Chain Monte Carlo QMCMC over traditional MCMC approach

     

             Portfolio Optimization

​

• With 16 qubits and quantum annealing, we optimize a portfolio and perform calculations faster than the classical approach.

     

             Stress Testing

​

• We utilize Generative Adversarial Networks  (GANs) and Variational AutoEncoders to generate synthetic data that retains the original data's statistical characteristics while generating new data points. This is particularly useful for creating economic scenarios during turbulent periods such as war, geopolitical tensions, and climate change. 

       

             Green AI

​

• Tensor networks in machine learning reduce the number of parameters in neural network models, lowering energy costs. 

​​

             Derivatives Pricing

​

• We showcase the superiority of Quantum Monte Carlo Simulation over classical Monte Carlo Simulation in terms of speed for pricing exotic options.