Analysis of Italian Word Embeddings

In our recent work, Analysis of Italian Word Embeddings (arXiv preprint arXiv:1707.08783), I and Stefano analysed the effect of different hyper-perparameters on the performances of two word embeddings algorithms, skip-gram (SG) and continuous-bag-of-words (CBOW), on semantic and morphs-syntactic tasks.

 

Results as accuracy with different hyper-parameters (y axis) using the 3COSADD (left bar) and the 3COSMUL (right bar) formula. The green part of the bars indicates the accuracy on the morpho- syntactic task whereas the red one the accuracy on the semantic task. The + sign on each bar indicates the accuracy on the entire dataset. The upper row of the figure shows the results of the SG algorithm and the bottom row the results of CBOW. The last two bars of the SG plots indicates the results obtained using the vectors made available by (Berardi et al., 2015)

 

The pre-trained vectors with the highest accuracy can be downloaded from the following link:

Please cite: Rocco Tripodi, Stefano Li Pira, Analysis of Italian Word Embeddings, arXiv preprint arXiv:1707.08783, if you use our them in your application.

@article{tripodi2017analysis,
title={Analysis of Italian Word Embeddings},
author={Tripodi, Rocco and Li Pira, Stefano},
journal={arXiv preprint arXiv:1707.08783},
year={2017}
}

Leave a Reply

Your email address will not be published. Required fields are marked *