A Better Variant of Self-Critical Sequence Training

TL;DR

This paper introduces an improved variant of Self-Critical Sequence Training that enhances performance by modifying the baseline function in the REINFORCE algorithm without additional computational costs.

Contribution

It proposes a simple change in the baseline function of REINFORCE, leading to better sequence training performance.

Findings

Improved sequence training results over standard methods.

No extra computational cost required.

Enhanced performance with a simple baseline modification.

Abstract

In this work, we present a simple yet better variant of Self-Critical Sequence Training. We make a simple change in the choice of baseline function in REINFORCE algorithm. The new baseline can bring better performance with no extra cost, compared to the greedy decoding baseline.