protom

v1.0.3

Published

a year ago

Build your awesome project with mathematical prediction, qubits memory and 3D render.

Downloads

0High
0Medium
0Low

yarflam

math quantum network render

Protom

Abstract

Protom is mathematical base to work with modern concept such like Machine Learning, Deep Learning, Quantum machine, 3D render or Data Analyze. The goal is to teach one language for many applications with an identical process. Behind your jobs, Protom can optimize the algorithms with the same solutions: balance between CPU / GPU / RAM, compression and standardization of data, easy import / export and compatibility.

The core of Protom is developed in Javascript. You can use it in your web and/or desktop applications. The Gitlab deposit is open source, you can fork or contribute. I will ask you to create units tests and comment your code.

Prerequisites

You should have to install Node.js on your computer (Windows / Linux / Mac / ARM systems).

Download here

Install

You can clone the project and install the node modules:

git clone https://gitlab.com/Yarflam/protom.git &&\
cd protom &&\
npm install

How to use it?

I write some papers in the /docs/papers folder:

Protom Development: see all features, structures and data types.
Quantum programming: an overview of quantum mechanics to work with qubits

Explore

Tree view

core
- bayes
  - Bayes.js: first test to apply Bayes' theorem
- helpers
  - gentools.js: access to random tools
    - Lack of clarity -> refactoring is required
- quantum
  - Qubit: simple quantum unit.
  - Qureg: Quantum Register (aka Q-System or Q-Space) - manipulate entranglement, quantum gates and measure.
- struct
  - controllers
    - Activation.js: activation functions
      - { Machine Learning }
    - Caller.js
      - Role: call a specific function in data type classes
      - Example: convert to another type
    - Operation.js
      - Role: perform a mathematical operation
    - Optimizer.js: use loss functions to minimize the errors
      - Example: Gradient descent
      - { Machine Learning }
  - datatypes
    - Complex.js: define a complex number
      - Example: 4 + 2i
    - Matrix.js: define a matrix
      - Notice: division is not supported
      - Example: [ 2, 5, 11 ]
    - Num.js: define a number
      - Example: 133.7
    - Quaternion.js: define a quaternion
      - Example: 20 + 21i + 7j + 8k
    - Tensor.js: define a tensor
      - Example: [ <vect1>, <vect2> ... <vectN> ]
    - Vector.js: define a vector
      - Example: < 3.1, 4.15, 1, 0, 1 >
  - nodes
    - Const: constants
    - Placeholder: dynamic values
    - Var: variables
- Protom.js: main application.
  - Role: manage the structures and orchestrates the operations
- index.js & web.js: different entry points (Node.js or web front)
docs
- draft
  - zun3D: 3D engine
- papers
  - DevelopmentPaper.ipynb: general
  - Quantum Programming.ipynb: quantum mechanics application
public: front pages (web)
sandbox:
- dev-playground.js: my tests on the Protom core (Yarflam)
- search-bayes: Test in application of Bayes' theorem
- search-map.js & search-map2.js: generate maps in command line (just for fun)
test: automated tests with Jest
- protom-func.spec.js: check the Protom core
- struct-datatypes.spec.js: check the data types (num, complex, quaternion, vector, matrix, tensor)
- struct-nodes.spec.js: check the nodes (const, var, placeholder)

Data types

| Type | About | Refers to | Example | | :--------- | :-------------------------------- | :--------------------------------------: | :--------------------: | | Num | Simple number | | 2.75 | | Complex | Complex number (real + imaginary) | Num | 4 + 2i | | Quaternion | Quaternion (w, i, j, k) | Num | 20 + 21i + 7j + 8k | | Vector | One dimension D | Num, Complex, Quaternion | <0.73, 0.12> | | Matrix | Two dimensions DxD | Num, Complex, Quaternion | [[ 1, 2 ], [ 3, 4 ]] | | Tensor | Multiple dimensions DxDx...D | Num, Complex, Quaternion, Vector, Matrix | [[ 1, 2 ], [ 3, 4 ]] |

Nodes

Controllers

Engines

Summary

Quantum System

In development...

Standards

Sequence of Q-bytes:

All gates are written in big-endian - an easy-to-use common representation - but they are executed in little-endian.

Compatibility with numbers:

You can use Number (or Num object) and Complex number.

Gates

| Gates | Code | Params | Matrix | | :------------------------------- | :---: | :----: | :--------------------------------------------- | | PauliX / SigmaX | X | 1 | [[ 0, 1 ], [ 1, 0 ]] | | PauliY / SigmaY | Y | 1 | [[ 0, -i ], [ i, 0 ]] | | PauliZ / SigmaZ | Z | 1 | [[ 1, 0 ], [ 0, -1 ]] | | Hadamard | H | 1 | 1/sqrt(2) * [[ 1, 1 ], [ 1, -1 ]] | | SqrtNot | | 1 | [[ (1+i)/2, (1-i)/2 ], [ (1-i)/2, (1+i)/2 ]] | | Phase | P | 1 + θ | [[ 1, 0 ], [ 0, e^(θi) ]] | | S-Gate | S | 1 | [[ 1, 0 ], [ 0, e^(π/2i) ]] | | T-Gate | T | 1 | [[ 1, 0 ], [ 0, e^(π/4i) ]] | | I-Gate / Identity | I | 1 | [[ 1, 0 ], [ 0, 1 ]] | | CNot | CN | 2 | see below | | CtrlZ | CZ | 2 | see below | | Swap | SW | 2 | see below | | SqrtSwap | RS | 2 | see below | | Toffoli | CCX | 3 | see below | | Fredkin | CSWAP | 3 | see below |

CNot

[[ 1, 0, 0, 0 ],
 [ 0, 1, 0, 0 ],
 [ 0, 0, 0, 1 ],
 [ 0, 0, 1, 0 ]]

CtrlZ

[[ 1, 0, 0,  0 ],
 [ 0, 1, 0,  0 ],
 [ 0, 0, 1,  0 ],
 [ 0, 0, 0, -1 ]]

Swap

[[ 1, 0, 0, 0 ],
 [ 0, 0, 1, 0 ],
 [ 0, 1, 0, 0 ],
 [ 0, 0, 0, 1 ]]

SqrtSwap

[[ 1,       0,       0, 0 ],
 [ 0, (1+i)/2, (1-i)/2, 0 ],
 [ 0, (1-i)/2, (1+i)/2, 0 ],
 [ 0,       0,       0, 1 ]]

Toffoli

[[ 1, 0, 0, 0, 0, 0, 0, 0 ],
 [ 0, 1, 0, 0, 0, 0, 0, 0 ],
 [ 0, 0, 1, 0, 0, 0, 0, 0 ],
 [ 0, 0, 0, 1, 0, 0, 0, 0 ],
 [ 0, 0, 0, 0, 1, 0, 0, 0 ],
 [ 0, 0, 0, 0, 0, 1, 0, 0 ],
 [ 0, 0, 0, 0, 0, 0, 0, 1 ],
 [ 0, 0, 0, 0, 0, 0, 1, 0 ]]

Fredkin

[[ 1, 0, 0, 0, 0, 0, 0, 0 ],
 [ 0, 1, 0, 0, 0, 0, 0, 0 ],
 [ 0, 0, 1, 0, 0, 0, 0, 0 ],
 [ 0, 0, 0, 1, 0, 0, 0, 0 ],
 [ 0, 0, 0, 0, 1, 0, 0, 0 ],
 [ 0, 0, 0, 0, 0, 0, 1, 0 ],
 [ 0, 0, 0, 0, 0, 1, 0, 0 ],
 [ 0, 0, 0, 0, 0, 0, 0, 1 ]]

Authors

Yarflam - initial work

License

The project is licensed under Creative Commons (BY-NC-SA) - see the LICENSE.md file for details.