Skip to content

Latest commit

 

History

History
41 lines (32 loc) · 1.79 KB

README.md

File metadata and controls

41 lines (32 loc) · 1.79 KB
Raw

Experiments towards adding logical constraints to multi-class, multi-label classifiers. This is the code for the paper "Class disjointness constraints as specific objective functions in neural network classifiers" published at SemDeep (http://semdeep.iiia.csic.es/SemDeep-2/index.html).

Prerequisites

See requirements.txt. This code is using the following libraries:

  • Tensorflow
  • Keras

How to run

  • Create the model_image_path and results_path directories specified in config.py (see below).
  • pip install -r requirements.txt
  • python train_both.py

Parameters can then be set in config.py (see below)

config.py

  • test_size: size of the test set wrt the dataset size
  • dataset_size: size of the dataset
  • nb_disjoint_classes: number of disjoint classes
  • nb_other_classes: number of other classes
  • epochs: number of training epochs
  • batch_size: batch size
  • dataset_seed: seed for the dataset random generation. Same seed will generate the same dataset.
  • disjoint_classes_output_weight: weight of the disjoint classes loss.
  • logdir_path: path for tensorboard logging. Logging can be disabled by removing the callback in the model training method.
  • results_path: path to store the trained models

Citations

If you find this code or the ideas useful, please reference the paper:

Class disjointness constraints as specific objective functions in neural network classifiers. François Scharffe. 2nd Workshop on Semantic Deep Learning (SemDeep). Montpellier, France. 2017.

Disclaimer

This is experimental code, meant to accompany the above referenced paper, not meant to use in production systems.

Copyright

Copyright 2017 Axon Enterprise, Inc.

License

See LICENSE.txt in this repository. Apache (version 2.0) https://www.apache.org/licenses/LICENSE-2.0