A Better Variant of Self-Critical Sequence Training

By Ruotian Luo. 2020

43 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Reinforcement Learning, Natural Language Generation

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Payel Das
2. Pierre L. Dognin
3. Igor Melnyk
4. Inkit Padhi
5. Yonatan Belinkov
6. Nadir Durrani
7. Hassan Sajjad
8. Fahim Dalvi

In Computer Vision and Pattern Recognition

1. Payel Das
2. Pierre L. Dognin
3. Igor Melnyk
4. Inkit Padhi

In Reinforcement Learning

1. Payel Das
2. Pierre L. Dognin
3. Igor Melnyk
4. Inkit Padhi

In Natural Language Generation

1. Payel Das
2. Pierre L. Dognin
3. Igor Melnyk
4. Inkit Padhi

Citations in the corpus

• Semantic Communications for Future Internet: Fundamentals, Applications, and Challenges (2022, unknown): “the long-term rewards it experiences, rather than to focus on estimating the reward, or how the reward function should be normalized [100], [101].”
• ReGen: Reinforcement Learning for Text and Knowledge Base Generation using Pretrained Language Models (2021, methodology): “Variations around REINFORCE exist on how to apply the gradients, such as MIXER (Ranzato et al., 2016), or on how to evaluate the baseline (Luo, 2020) to minimize the gradient variance.”