Study guide

A route through the course

1. Quantum Algorithmic Thinking: Learn the design process that the whole course follows: state the problem and access, encode the property, engineer interference, construct the readout, prove correctness, and account for resources.
2. Coherent Computation: Build clean coherent subroutines: function evaluations and arithmetic that leave exactly the requested output, restore their workspace, and offer the inverse and controlled access later algorithms need.
3. Phase Encoding: Encode information in relative phases and read it with interference: eigenstate kickback, the Hadamard test, sign patterns on many qubits, hidden-structure sampling, and finer phases from computed values.
4. Fourier Methods: Build the quantum Fourier transform and use it with controlled powers and kickback to estimate eigenphases, find orders, and factor integers.
5. Amplitude Manipulation: Raise small success probabilities by composing reflections, control the rotation with phase sequences, estimate probabilities with amplitude estimation, and know the limits of these speedups.
6. Quantum Walks: Replace random walks by coherent walks, analyze them through their spectra, and use the square-root relation between classical spectral gaps and quantum phase gaps to build search algorithms.
7. Operator Embedding: Represent matrices that are not unitary as blocks of unitaries, build such block encodings from sums of unitaries, compose them, and account for subnormalization, success probability, and access costs.
8. Polynomial Transformations: Apply polynomial functions to the eigenvalues or singular values of a block-encoded matrix, and design algorithms for filtering, inversion, and simulation by choosing polynomials.
9. Evolution Engineering: Engineer time evolution: build e^{−iHt} from short evolutions of simple terms with controlled error, compare simulation strategies fairly, and use slowly changing Hamiltonians to prepare ground states.
10. Variational Methods: Understand variational algorithms as heuristics: parameterized circuits, measured objectives, classical optimization, and the evidence needed to judge their performance.
11. Adaptive Measurement: Use measurement inside a computation: branch on outcomes, defer or advance measurements, estimate phases adaptively with one qubit, and compute by measurement with teleported gates and injected magic states.

Dependencies between techniques

Coherent computation (Chapter 2) supplies the clean subroutines that every later oracle assumes. Phase encoding (Chapter 3) and the Fourier transform (Chapter 4) combine into phase estimation, which Chapter 5 reuses for amplitude estimation and Chapter 11 rebuilds with one control qubit and feedback. Reflections from Chapter 5 reappear as the steps of quantum walks in Chapter 6 and as the walk operators of qubitization in Chapter 8. Block encodings (Chapter 7) are the input to polynomial transformations (Chapter 8), which give one of the two simulation strategies compared in Chapter 9. Variational methods (Chapter 10) trade the guarantees of the earlier chapters for shallow circuits and classical optimization, and say so.

Working through a subsection

Read the explanation once, then redo its central calculation without looking. When a subsection asks you to predict, write the prediction and your reason before opening the answer. The three problems at the end of each subsection move from checking the idea, to applying it in a new instance, to exposing a limit or a common incorrect implementation. Attempt them in order. The chapter exercises at the end combine several subsections; open hints one at a time.

Assumptions to keep visible

A claimed advantage must state the input model, the output required, the allowed error, and the resources counted. Queries, gates, depth, qubits, precision, repetitions, postselection, and classical processing are different resources. Access to a unitary does not automatically include its inverse or a controlled version, and a quantum state encoding a vector is not a classical list of its entries. Small worked circuits check the mathematics of an example; they do not establish scaling.

Checking the worked examples

The repository includes numerical checks of the small instances used in the text (education/algorithms/checks/worked_examples.py and education/checks/check_mathematics.py). They confirm amplitudes, probabilities, and identities for the stated examples; they do not prove the general theorems the text cites.

Reader controls

Use Contents to move between topics, Search (or the / key) to find an idea, and A− or A+ to adjust the text. Print chapter expands hidden hints and solutions for a printed copy. Your text size and last reading position are kept only in this browser.