How to embed lexical features

How to embed lexical features

What are the lexical features you consider?

e.g., Embedding Lexical Features via Low-Rank Tensors

I want to embed the part of speech feature and person feature into the word vector

The part of speech feature and the person feature can be encoded into one-hot vectors x and y, respectively. We can get the embedded vector v from x and y:

z = concat([x, y])
v = z * E

where E is a trainable embedding matrix.