Efficient Estimation of Word Representations in Vector Space
We propose two novel model architectures for computing continuous vector representations of words from very large data sets. The quality of these representations is measured in ...
We propose two novel model architectures for computing continuous vector representations of words from very large data sets. The quality of these representations is measured in ...
In recent years, probabilistic features became an integral part of state-of-the-are LVCSR systems. In this work, we are exploring the possibility of obtaining the features direc...