I continued my research about representing the word2vec algorithm’s by a spatial visualisation. It tries to answer the simple question: what is a “dog” (chien) for the machine. The expriment shows different intepretation of this word into different models trained on different texts. It allows to see surrounding words .
The experimentation shows three word2vec corpus :
- Verlaine’s Poème Saturniens
- Lorris’ Romand de la Rose
Each text are shown with different font and color.
The experiment work on smarphone at the address: