Quantum Computing Use Cases
Market landscape
Market segments: hardware, QPU provider, software offering, applications, users.
Use cases
Walks through a survey on quantum computing in enterprise: top 3 challenges of enterprise adopting quantum computing: expense/business case justification, skills/lack of internal competency, don’t see where/how it could be used; business value: increased efficiency, profitability, improved processes/productivity/revenue; use cases are all optimization problems (what variables/weights to form the best solution): resource scheduling, financial forecasting/risk management, logistics management, AI/ML, materials science, fluid dynamics, chemical engineering, pharmaceutical development.
Financial
fraud detection, portofolio optimization, option pricing, clustering, scenario analysis (many scenarios)
PQC/post-quantum cryptography
D-wave mentions primary approach of annealing (suited for combinatorial optimization) vs. gate-model (differential equation). 51% financial companies want to solve optimization problems, e.g. Risk analysis, Portfolio optimization, Fraud detection, Product pricing and insurance underwriting, AI/ML.
Case study: caixabank portofolio optimization, mastercard fraud detection, goodlabs real time quantum liquidity optimizer, satispay quantum-optimized rewards program (uses TF wide-deep model to predict CTR/click-through rate, CQM hybrid solver time per simulation 3h from week)
Interview of David Isaac, co-founder of quantum finance company Abaqus
JP Morgan Chase talk on QWC 2023. Financial market is volatile, speed is important
Mastercard sees near-term success with annealing, but gate-model has long-term potential. Use cases: credit card offer allocation, hidden flow discovery/anti-money laundering (crypto-mixing services), fraud detection, cross-border, multi-lateral net settlement process (with demo)
Challenging the security of current encryption, optimization (Quantum Approximate Optimization Algorithm (QAOA) for real-time portfolio adjustments, Quantum Monte Carlo (stock market simulation, portfolio evaluation, and derivatives pricing), High-dimensional Optimization (Credit Scoring/Multiverse, Option Pricing and Derivatives/JPMorgan))
Several quantum algorithms show promise in enhancing fraud detection capabilities:
- Quantum Support Vector Machines (QSVM): These can classify transactions as legitimate or fraudulent with higher accuracy by finding optimal hyperplanes in high-dimensional spaces.
- Quantum Neural Networks (QNN): By mimicking the structure of classical neural networks, QNNs can learn complex patterns in transaction data, improving the detection of subtle fraud indicators.
Chemistry
molecule simulation
drug discovery
ground state energy/GSE, excited state energy/ESE, Born-Oppenheimer Potential Energy Surface/BOPES
Compute how molecules react with each other, energy and how atoms move around in chemical reaction, from a catalyst, which needs accuracy. Quantum computing can mimic how nature computes.
Cybersecurity
Post-quantum cryptography, quantum-resistant encryption: Lattice-based cryptography (LBC)
Machine Learning
Talks about in ML classification, kernel function is used to map data points into higher-dimension feature space for easier classification. Kernel function can explode as data complexity increases. Quantum computers can access much more complex and higher features spaces, offering exponential speed-up on certain classification problems.
Appendix
Classic computers use transistors as the physical building blocks of logic. In quantum computers they may use trapped ions, superconducting loops, quantum dots or vacancies in a diamond.
Quantum computing milestones
