| name | cv-photonic-qnn-edge-medical |
| description | Parameter-efficient Continuous-Variable Photonic Quantum Neural Networks (CV-QNN) for edge medical AI applications. Covers room-temperature quantum computing on photonic hardware, MobileNet feature extraction + PCA dimensionality reduction, and CV-QNN classifiers for medical image classification (oral cancer, dermatology, radiology). Use when: (1) building edge-deployable quantum ML for healthcare, (2) comparing qubit vs CV photonic approaches, (3) designing parameter-efficient quantum classifiers for resource-constrained medical settings, (4) implementing hybrid classical-CV quantum pipelines for medical image classification. Trigger words: cv-qnn, photonic quantum, continuous-variable, edge quantum AI, oral cancer detection, parameter-efficient quantum, room-temperature quantum, mobile medical AI, smartphone screening.
|
CV Photonic QNN for Edge Medical AI
Based on arXiv:2606.28252 — "Parameter-Efficient Continuous-Variable Photonic Quantum Neural Networks for Edge Quantum AI: Demonstration in Oral Cancer Detection"
Core Architecture
MobileNetV1 feature extractor → PCA (16-dim) → CV-QNN classifier (4 qumodes, 2 layers)
Key Design Principles
- Room-temperature operation: CV photonic QCs operate without cryogenics, enabling edge deployment
- Parameter efficiency: 4-qumode CV-QNN uses ~100× fewer trainable parameters than equivalent qubit VQC
- Dimensionality matching: PCA to N dimensions maps to N qumodes — keep ≤16 for near-term hardware
- Hybrid classical-quantum: Classical backbone (MobileNet/ResNet) + quantum classifier head
Pipeline Steps
- Feature extraction: Pretrained MobileNetV1 (or ResNet18) extracts 512-dim features from medical images
- Dimensionality reduction: PCA to ≤16 dimensions to match available qumodes
- CV-QNN encoding: Amplitude encoding of PCA features into qumode states
- Quantum layers: 2 layers of displacement + rotation + Kerr gates per qumode
- Measurement: Homodyne detection → classical post-processing → classification
Hardware Requirements
- CV photonic quantum processor (e.g., Xanadu X8, Orquestra)
- Room-temperature operation (no cryogenics)
- Classical edge device (smartphone, Raspberry Pi) for feature extraction
Comparison: Qubit vs CV Photonic
| Aspect | Qubit VQC | CV Photonic QNN |
|---|
| Operating temp | ~15mK (cryogenic) | Room temperature |
| Parameters per layer | 2^n scaling | Linear in qumodes |
| Edge deployable | No | Yes |
| Measurement | Projective | Homodyne/heterodyne |
| Encoding | Amplitude/angle | Displacement/squeezing |
Workflow
When to Use
- Medical image classification on edge devices with limited compute
- Low-resource clinical settings without cryogenic infrastructure
- Smartphone-based screening applications
- When parameter efficiency is critical (embedded systems)
When NOT to Use
- Tasks requiring deep quantum circuits (>10 layers)
- Problems needing large Hilbert space (>20 qumodes)
- High-precision tasks where classical models already saturate
Activation Keywords
cv-qnn, photonic quantum, continuous-variable, edge quantum AI, oral cancer detection, parameter-efficient quantum, room-temperature quantum, mobile medical AI, smartphone screening, quantum edge computing, quantum medical classification
