Vec2Sent: Probing Sentence Embeddings with Natural Language Generation

11/01/2020
by   Martin Kerscher, et al.
0

We introspect black-box sentence embeddings by conditionally generating from them with the objective to retrieve the underlying discrete sentence. We perceive of this as a new unsupervised probing task and show that it correlates well with downstream task performance. We also illustrate how the language generated from different encoders differs. We apply our approach to generate sentence analogies from sentence embeddings.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset