Open-AML-Engine

An open Algebraic Machine Learning (AML) engine for building models and performing inference. Designed for community-driven exploration and research in AML.

The current version has been tested on Linux.

Installation

1. Create a new Python environment (recommended)

2. Install dependencies:

• cffi
• networkx (only for example03)
• matplotlib (only for example03)

3. Compile C-extensions

cd aml/aml_fast
python build_amlFastLibrary.py

4. Install AML library.

From the root folder, where setup.py is placed, run the following command:

python -m pip install -e .

Note that the command ends with a dot.

5. Documentation and demos

Detailed information and tutorials on the engine can be found at the AML Engine API webpage.

Interactive demos also available on Google Colab:

• Vertical Bars Demo
• MNIST Demo
• Hamiltonian Graphs Demo

6. Extra

If you plan to run example04_mnist.py you will need to obtain the MNIST dataset and place the files in Examples/mnist_datasets (the path can be modified in MNIST.py):

• t10k-images-idx3-ubyte
• t10k-labels-idx1-ubyte
• train-images-idx3-ubyte
• train-labels-idx1-ubyte

The dataset can be downloaded from Yann Lecun MNIST database or Huggingface.

How to cite

Work done with this engine should be cited with Algebraic Machine Learning: Learning as computing an algebraic decomposition of a task.

FAQ

When using the library I get "Illegal instruction (core dumped)"

It is possible that AVX or SSE are not supported. If that is the case, remove the following lines from build_amlFastLibrary.py.

"-msse2",
"-march=core-avx2",

---

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 952091.