Week 23 - my notes

Week 23 - VQE and its Applications

Understanding how a molecule is composed and structured, you can use this knowledge to explain its function and properties
We can reverse engineer this process, by creating something that you want to behave in a certain way
The problem is that finding a structure is hard. Finding its composition is easy
Quantum computers will help us find these structures!
Bigger molecule = more calculation = more computing power

As the size of the protein increases, the number of possible solutions for its structure rises exponentially
One method of developing vaccines is to disrupt the process of a virus’ spike protein from binding to human cells
In order to find a molecule that can stop this process, we need to try many simulations
In pharmaceuticals, you’re dealing with large proteins with thousands of atoms
To experimentally determine a protein structure, you use x-ray or electron diffraction, Nuclear Magnetic Resonance, and Cyro electron microscopy
To computationall determine a protein structure, you use molecular dynamics simulations on a supercomputer, crowdsourced simulations (folding@home), and artificial intelligence
Why to use quantum? Quantum systems are naturally discrete, so we can map the different possible solutions to the optimization problem to discrete quantum levels
Because of current hardware errors, results aren’t reliable, therefore we use VQE to shift some of the work to classical computers

CPU receives atoms in the protein: We know what atoms are present, just not how they are arranged three-dimensionally
CPU decides which structure to try out and sends to QPU: The structure will get implemented using a tunable quantum circuit. A lot of math is used here.
QPU measures the energy of the structure: This is very similar to the circuit measurements we made in the labs
QPU sends results of the measurement to CPU:
CPU decides if this structure minimizes energy - if yes, solved. If not, repeat the process

The world of finance is built upon massive data sets
Quantum applications in finance are already in motion
QC could solve High-Frequency Trading, Risk Profiling, Portfolio Optimization
Given a choice of available assets (cash, stocks, commodities), what is the best combination of assets to maximize your returns and minimize risk?
CPU receives the assests in the portfolio
CPU decides which portfolio to try and sends to QPU
QPU measures the expected value of the portfolio
QPU sends results of the measurement to CPU
CPU decides if this portfolio maximizes returns - if yes, solved. If not, repeat the process

Quantum protocols: Sending and receiving qubits for cryptography and efficiency communication (E.g. Superdense coding)
Longer-term algorithms: Powerful algorithms that assume perfect, fault-tolerant hardware with millions of qubits, which is likely decades away (Grover’s algorithm)
Near-term algorithms: Leverage quantum computers alongside classical computers to perform functions more efficiently than classical alone (VQE)