Embarrassingly Simple Zero Shot Implementation

This branch includes an implementation in python of the paper:

Romera-Paredes, Bernardino, and P. H. S. Torr. "An embarrassingly simple approach to zero-shot learning." Proceedings of The 32nd International Conference on Machine Learning. 2015.

http://jmlr.org/proceedings/papers/v37/romera-paredes15.pdf

The original Matlab code and data is available here: https://github.com/bernard24/Embarrassingly-simple-ZSL https://dl.dropboxusercontent.com/u/5961057/ESZSL_v0.1.zip

Note

I borrowed from the idea in https://github.com/MLWave/extremely-simple-one-shot-learning by using 2 components from PCA and LLA (Locally Linear Embedding) as features to classify the digits dataset from sklearn. Overall it uses the same exact approach as the paper to calculate V and W.