Quantum Computing To My Grandmother

I've been mulling over quantum computing for months and why it's going to be so groundbreaking. Today, reading this article from FastCompany,

I wanted to learn why the designs of solutions for a classical computing architecture and a quantum computing architecture are different.

And ChatGPT has helped me understand it. Look at this:

Me: Show me an example of designing a solution for a quantum computer and the same solution for a classical computer.

ChatGPT4: Sure, let's consider a common problem in both classical and quantum computing: optimizing a complex function.

For simplicity, let's use the task of finding the minimum of a given function, a problem commonly encountered in various fields like finance, logistics, and machine learning.

Classical Computing Solution: Gradient Descent

Problem: Find the minimum of a function f(x).

Solution: Use the Gradient Descent algorithm.

1. Initialize: Start with a random point x0.

2. Iterate:

• Compute the gradient (derivative) of f at the current point.
• Update the point by moving in the direction opposite to the gradient.

x_new = x_old - eta x gradient of f(x_old)        

where eta is the learning rate.

3. Convergence: Repeat until the changes in x are below a certain threshold.

This method iteratively adjusts the input to minimize the function, leveraging calculus to navigate the function's landscape.

Quantum Computing Solution: Quantum Annealing or Variational Quantum Eigensolver (VQE)

Problem: The same as above, but we leverage quantum properties for optimization.

1. Quantum Annealing:

• Encode the optimization problem into a quantum system.
• Start with a superposition of all possible states.
• Gradually evolve the system using quantum annealing, which guides the system towards its lowest energy state, corresponding to the minimum of the function.

2. Variational Quantum Eigensolver (VQE):

• Encode the problem into a quantum circuit.
• Use a quantum computer to prepare a quantum state (represented by qubits).
• Measure the state to evaluate the function.
• Use a classical optimizer to adjust the parameters of the quantum circuit, aiming to minimize the function.
• Iterate this process, using the quantum computer for state preparation and measurement, and the classical computer for optimization.

Both quantum methods exploit quantum phenomena like superposition and entanglement. Quantum Annealing is particularly powerful for problems with rugged energy landscapes, while VQE is more versatile and can be adapted to various optimization problems.

Key Differences:

• Classical Solution: Uses a deterministic, step-by-step approach. Suitable for a wide range of problems but can struggle with highly complex or multi-dimensional landscapes.
• Quantum Solution: Leverages quantum properties to explore multiple possibilities simultaneously. Can potentially find solutions faster for specific types of problems (like those with many local minima), but requires a quantum computer and is more experimental in nature.

This example illustrates how the same problem can be approached differently in classical and quantum computing, each with its own strengths and limitations.

My perception

I recently had the opportunity to explain to a group of clinical laboratory experts the difference between designing a classic algorithmic solution and doing so using machine learning algorithms.

One of the main messages was that classic algorithmic presented serious difficulties in responding to complex problems and that, there, AI or similar does it much better.

Reading this comparison between quantum and classic solutions, I get the feeling that we will be able to solve very complex problems again without the need to access as much data as AI-based solutions require, but only with a "process model".

How do you feel about that? :)