Neural Sequence Model Training via α-divergence Minimization

Code

• github.com/sotetsuk/alpha-dimt-icmlws
  OfficialIn paperpytorch★ 0

Abstract

We propose a new neural sequence model training method in which the objective function is defined by -divergence. We demonstrate that the objective function generalizes the maximum-likelihood (ML)-based and reinforcement learning (RL)-based objective functions as special cases (i.e., ML corresponds to 0 and RL to 1). We also show that the gradient of the objective function can be considered a mixture of ML- and RL-based objective gradients. The experimental results of a machine translation task show that minimizing the objective function with > 0 outperforms 0, which corresponds to ML-based methods.

Tasks

Reproductions