Algorithm Library Guide

Overview

This chapter covers browsing, filtering, and viewing details in the algorithm library. It fits these scenarios:

• You want to see which algorithms the platform offers before submitting a task.
• You already know the rough direction but need categories or keywords to narrow the scope.
• You are on an algorithm detail page and want to quickly decide whether to read the description, tune parameters, or check results next.

Topics: Browse the Library, Search & Filter, View Algorithm Details, Supported Algorithms.

Page Overview

The algorithm library page has the following main sections:

Numbers in the figure correspond to:

1. Left navigation: switch among home, quantum programming, the algorithm library, and the PDE solver.
2. Filter & search bar: the category dropdown on the left and the keyword search input on the right, used to narrow down the browse scope.
3. Algorithm card area: shows the currently available algorithms; click any card to enter the detail page.
4. AI chat panel: a persistent right-side panel for asking questions about algorithm theory, parameters, or task results; see the AI Chat Guide for details.

When the list is loading, failed to load, or has nothing to display, the page shows the matching state prompt first.

Algorithm List

The algorithm list lives in the main content area. It is used to browse the currently available algorithms; click an item to enter the detail page.

• Click an algorithm item to enter details.
• The list content is driven by the page’s internal state, and may display loading, error, or empty states.
• On a load failure, the page supports retry.

Search & Filter

The filter bar narrows the browse scope and contains:

• Category tabs
• Keyword search

Recommended order:

1. Browse the default list first to get a sense of the available algorithms.
2. Switch the category tab when the scope is too wide.
3. Use keyword search once your target is clear.

Algorithm Details

Clicking an algorithm card opens the algorithm detail page:

Numbers in the figure correspond to:

1. Top action bar: back-to-library on the left, algorithm name in the middle, run button on the right.
2. Algorithm description area: shows the algorithm’s principle, run modes, and parameter docs; content comes from the algorithm’s README.
3. Parameter panel: configure run parameters such as the number of qubits and the run mode.
4. Task result area: view task information, the quantum circuit, and computation results.
5. Run button: after filling in the parameters, click to submit the algorithm task.
6. AI chat panel: a persistent right-side panel for asking questions about algorithm theory, parameters, or task results; see the AI Chat Guide for details.

To run an algorithm:

1. Read the parameter meanings in the algorithm description area (②) to understand each field’s purpose and value range.
2. Adjust the defaults in the parameter panel (③) if needed.
3. Click the run button (⑤) to submit the task.
4. Switch tabs in the task result area (④) to view task information, the quantum circuit, and computation results.

Suggestions

• The algorithm library is meant for browsing first and diving in later — you do not need to open every item on the first visit.
• To quickly confirm whether a specific algorithm exists: use the filter and search first.
• If an algorithm fits what you want to do: stay on the detail page instead of bouncing back to the list repeatedly.

Supported Algorithms

This section provides only a concise overview of the algorithms available in the algorithm library. For detailed principles, parameters, examples, and API usage, refer to the UnitaryLab Algorithms User Manual.

Variational Quantum Linear Solver (VQLS)

This algorithm solves structured linear systems with a variational ansatz. It defines a local Hadamard-test-based cost function and optimizes circuit parameters with COBYLA, then reports fidelity, relative error, residual norm, solution states, and generated circuit artifacts.

Variational Quantum Eigensolver (VQE)

This algorithm estimates the ground-state energy of a transverse-field Ising Hamiltonian. It builds a parameterized circuit with Ry and CX gates, minimizes energy expectation with COBYLA, evaluates spin observables, and outputs energy metrics, runtime, circuit, convergence, and observable plots.

Amplitude Amplification

The algorithm applies the amplitude amplification method by preparing an initial state and repeatedly executing Grover iterations with an Oracle and a diffuser to boost the target state measurement probability on a quantum circuit.

Quantum Amplitude Estimation (QAE)

The algorithm applies quantum amplitude estimation by constructing a Grover operator and performing phase estimation with IQFT to extract the phase bit-string and recover the target state amplitude for numerical approximation.

Cartan Decomposition

The algorithm takes a Hamiltonian and evolution time to build an approximate evolution circuit via the Cartan-Lax flow method, comparing the result with the exact evolution matrix to simulate complex dynamics.

Continuous Variable Quantum Neural Network (CVQNN)

The algorithm builds a continuous-variable quantum neural network in truncated Fock space using displacement, squeezing, and Kerr nonlinearity, optimized via Adam for binary classification tasks on two-dimensional input datasets.

Discrete Logarithm

The algorithm solves the discrete logarithm problem by encoding period information through Quantum Phase Estimation, combining classical continued fractions and modular arithmetic to efficiently recover exponents in modular power congruence expressions.

Grover Search

The algorithm prepares a uniform superposition and applies repeated Grover iterations consisting of an Oracle and a diffuser to amplify target amplitudes, enabling high-probability searching of specific bit-strings in databases.

Hadamard Test

The algorithm uses an ancilla qubit and controlled unitaries to estimate the real or imaginary parts of an operator’s expectation value, state overlap fidelity, or complex phases through statistical sampling.

Hadamard Transform

The algorithm applies Hadamard gates to prepare a uniform superposition from initial states or verifies the self-inverse property by applying the transform twice, serving as a fundamental building block for parallelism.

HHL for Linear Systems

The algorithm solves Hermitian linear systems by integrating Quantum Phase Estimation, controlled reciprocal rotations, and inverse QPE to estimate solution vectors with potential exponential speedup over classical methods.

Linear Combination of Unitaries (LCU)

The algorithm represents non-unitary operators as a linear combination of unitaries using ancilla-state preparation and controlled selection to probabilistically apply the target operator on a system quantum register.

Quantum Approximate Optimization Algorithm (QAOA)

The algorithm solves Max-Cut problems by alternating cost and mixer layers in a parameterized circuit, using the COBYLA optimizer to find optimal angles that yield high-probability bit-strings for graph partitioning.

Quantum Circuit Born Machine (QCBM)

The algorithm uses a Quantum Circuit Born Machine with parameterized rotations and entanglers, trained via Adam and the parameter-shift rule to learn discrete target distributions by minimizing KL divergence.

qDrift Algorithm

The algorithm approximates time evolution by randomly sampling Pauli terms from a decomposed Hamiltonian, building a random product formula that simulates quantum dynamics with tunable step counts and circuit depth.

Quantum Fourier Transform (QFT)

The algorithm implements discrete Fourier transforms and their inverses on quantum registers, transforming state amplitudes into the frequency domain through rotation and controlled gates for use in advanced algorithms.

Quantum Phase Estimation (QPE)

The algorithm extracts the phase of a unitary operator relative to its eigenstate by using controlled power operations and IQFT to translate phase information into a decimal register estimate.

Quantum Signal Processing (QSP)

The algorithm performs quantum signal processing on a single qubit using alternating signal and phase operators, searching for optimized phase sequences to obtain high-precision polynomial approximations of target functions.

QSP-based Hamiltonian Simulation

The algorithm simulates Hamiltonian evolution by block-encoding matrices and using QSP to approximate sine and cosine components, combining them via LCU to build precise time-slice evolution circuits.

QSVT-Based Linear-System Solver

The algorithm utilizes Quantum Singular Value Transformation to solve linear systems by block-encoding the coefficient matrix and applying polynomial transformations to compute solution vectors with high accuracy.

Shor’s Factoring Algorithm

The algorithm factors composite numbers by using quantum phase estimation to find periodicity in modular exponentiation, combined with classical fraction methods to identify prime factors in polynomial time.

Simon’s Algorithm

The algorithm finds a hidden binary mask $s$ for a black-box function using quantum superposition to generate linear equations, which are then solved via classical Gaussian elimination.

Taylor-Series Hamiltonian Simulation

The algorithm implements Hamiltonian evolution by truncating Taylor series and using LCU on a Pauli-decomposed basis to build an approximation circuit that satisfies target error constraints.

Suzuki-Trotter Product Formula

The algorithm uses the Suzuki-Trotter product formula to approximate time evolution by decomposing the Hamiltonian into non-commuting terms, supporting higher-order constructions through repeated time slices.

Variational Quantum Classifier (VQC)

The algorithm maps classical data into quantum states and uses parameterized circuits trained via Adam and the parameter-shift rule to perform supervised classification on datasets like Iris.