Autoregressive Text Generation beyond Feedback Loops

Florian Schmidt
Florian Schmidt

EMNLP/IJCNLP (1), pp. 3398-3404, 2019.

Cited by: 3|Views42
EI

Abstract:

Autoregressive state transitions, where predictions are conditioned on past predictions, are the predominant choice for both deterministic and stochastic sequential models. However, autoregressive feedback exposes the evolution of the hidden state trajectory to potential biases from well-known train-test discrepancies. In this paper, we...More

Code:

Data:

Your rating :
0

 

Tags
Comments