We have made available a vector word model learned with neural networks

The published model describes words of natural language by means of multidimensional vectors. The vectors refer to words’ hidden properties which are motivated by the layers of the neural network. The model may be used for determining the degree of word similarity to which specific vectors relate and thus for the generation of word frequency lists.

The model is available at our repository on the D-Space platform: https://clarin-pl.eu/dspace/handle/11321/327

Posted in Bez kategorii.