Revolutionizing Medical Imaging with Quantum Machine Learning
Classical machine learning struggles with the complexity and high dimensionality of medical imaging data. Machine learning models trained on specific datasets may struggle to generalize across different patient populations. This leads to slower, and potentially inaccurate diagnoses that can impact patient care. Ensuring the robustness and generalizability of machine learning models is crucial to their successful deployment in clinical practice.
Research
Lead Researchers: Laia Domingo & Mahdi Chehimi
Ingenii is developing an end-to-end quantum machine learning pipeline that can process high-dimensional data efficiently.
Our pipeline will include:
Medical imaging data (MRI or mammograms)
Quantum filter to enhance image quality
Ability to make predictions on tumor segmentation/classification
Benefits
Speed & Accuracy
Our QML pipeline aims to outperform classical ML solutions in speed and accuracy of medical imaging analysis.
Efficiency
Allows for the efficient handling of large and complex medical imaging datasets.
Enhanced Transfer Learning
Our quantum-enhanced transfer learning-based model will aim to enhance the area under the curve by more than 30%.
Enhanced Precision & Specificity
Our hybrid architecture will aim to achieve an enhancement of above 5% in both precision and specificity compared to state-of-the-art classical models.
Faster Training Time
Our quantum filter has the potential to make our model twice as fast as classical state-of-the-art models in terms of training time.
Reduction in GPU Resources
By leveraging tensor networks in our proposed pipeline, we aim to achieve 75-90% reduction in the required GPU resources to train the ML models.