L6-10 · Design a quantum causal experiment

Learn¶

Start with a precise question and a list of candidate mechanisms. State which systems enter and leave each laboratory. Separate a direct channel from a common source where the question requires that distinction. List the operations and observations actually available to the experimenter.

Calculate predictions before collecting data. Find settings where the candidates differ, then allocate trials with a declared rule. Include calibration, drift controls, and a method for uncertainty. Reserve independent data for checking a fitted model. Report failures as evidence against the model assumptions.

A good design can return an inconclusive result. Two mechanisms may remain equivalent under restricted operations. In that case, report the equivalence or a bound rather than forcing a single graph. A larger model family needs its own identifiability and statistical analysis.

Practice¶

Check your understanding¶

What is the correct result when all allowed predictions match?

• A. Report an unresolved equivalence under the stated design

• B. Choose the simpler graph as proven true

• C. Increase decimal precision until they differ

Sources¶

This lesson draws on Costa & Shrapnel, 2016Barrett et al., 2019Watrous, 2018.

Next: L7-01 · Define the learning task