Skip to content
No description, website, or topics provided.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
README.md
conv_cne.py
knn.py
naive_bayes_nearest_neighbours.py
nbknca.py
optimize.py
pca.py
test_ccml2_mnist.py
test_ccml2_norb.py
utils.py

README.md

Learning a Metric for Class-Conditional KNN

Python (Theano) implementation of Learning a Metric for Class-Conditional KNN code provided by Daniel Jiwoong Im and Graham W Taylor.

Class Conditiaonl Metri Learning (CCML) learn a metric which captures perceptual similarity. Similar to how Neighbourhood Components Analysis optimizes a differentiable form of KNN classification, which optimizes a soft form of the Naive Bayes Nearest Neighbour (NBNN) selection rule. For more information, see

@article{Im2016ccml,
    title={Learning a Metric for Class-Conditional KNN},
    author={Im, Daniel Jiwoong and Taylor, Graham W.},
    journal={International Joint Conference on Neural Networks (To appear)},
    year={2016}
}

If you use this in your research, we kindly ask that you cite the above workshop paper

Dependencies

Packages

How to run

Entry code for one-bit flip and factored minimum probability flow for mnist data are

    - /test_ccml2_mnist.py
    - /test_ccml2_norb.py
You can’t perform that action at this time.