Feed-forward

What is feed-forward?   

Feed-forward is a quantum processing technique in which measurement outcomes obtained during a computation are used in real time to dynamically update and control subsequent operations. Unlike fixed (static) quantum circuits, feed-forward enables adaptive operations, where the system is updating according to intermediate results. This capability is essential in photonic quantum computation as we will detail later.   

How does feed-forward work?

1. Intermediate quantum measurements: specific qubits (or photonic modes) are measured during the computation.  
2. Classical signal processing: the measurement results are very rapidly processed by classical electronics.  
3. Conditional quantum operations: depending on the classical signal, a given quantum operation is applied, and the system evolves through an updated unitary.   

Why is feed-forward crucial for photonic quantum computation?  

In schemes like Fusion-based quantum computation, feed-forward plays an essential role. Specifically, it is required for:   

• The fusion of smaller entangled states to create larger entangled “resource states”. 
• The routing of these resource states in computation.  
• Adaptively updating measurement based on prior measurement results.  

In schemes like Spin-optical quantum computation , feed-forward is necessary to implement Repeat-Until-Success gates.  

How to experimentally realize feed-forward?

To experimentally realize feed-forward, we need fast electronics (FPGA or ASIC), and ultra-high performance integrated platform (with lithium niobate LiNbO₃ or barium titanate BTO).  

Frequently asked questions about error-corrected qubits  

1. Why is feed-forward so important for photonic systems? Photons are great because they remain very coherent, but they also do not naturally interact, so many effective quantum gates rely on measurement outcomes and conditional operations.   
2. Can feed-forward help correct errors? Yes. Some error-correction schemes benefit from conditional operations based on measurement results. Adaptive strategies have been shown to increase the noise threshold for error correction.