Exposure Bias versus Self-Recovery: Are Distortions Really Incremental for Autoregressive Text Generation?

By Tianxing He, Jingzhao Zhang, Zhiming Zhou, James R. Glass. 2019

51 citations recorded by the source corpus.

Exposure bias has been regarded as a central problem for auto-regressive language models (LM). It claims that teacher forcing would cause the test-time generation to be incrementally distorted due to the training-generation discrepancy. Although a lot of algorithms have been proposed to avoid teacher forcing and therefore alleviate exposure bias, there is little work showing how serious the exposure bias problem actually is. In this work, we focus on the task of open-ended language generation, propose metrics to quantify the impact of exposure bias in the aspects of quality, diversity, and consistency. Our key intuition is that if we feed ground-truth data prefixes (instead of prefixes generated by the model itself) into the model and ask it to continue the generation, the performance should become much better because the training-generation discrepancy in the prefix is removed. Both automatic and human evaluations are conducted in our experiments. On the contrary to the popular belief in exposure bias, we find that the the distortion induced by the prefix discrepancy is limited, and does not seem to be incremental during the generation. Moreover, our analysis reveals an interesting self-recovery ability of the LM, which we hypothesize to be countering the harmful effects from exposure bias.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing

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. Pascale Fung
2. Andrea Madotto
3. Wenliang Dai
4. Nayeon Lee
5. Yan Xu
6. Etsuko Ishii
7. Ziwei Ji
8. Dan Su

In Computer Vision and Pattern Recognition

1. Pascale Fung
2. Wenliang Dai
3. Ziwei Ji
4. Dan Su
5. Nino Vieillard
6. Tiezheng Yu
7. Matthieu Geist
8. Olivier Bachem

In Natural Language Processing

1. Pascale Fung
2. Nayeon Lee
3. Andrea Madotto
4. Etsuko Ishii
5. Yejin Bang
6. Yan Xu
7. Nino Vieillard
8. Rita Frieske

Citations in the corpus

• Survey of Hallucination in Natural Language Generation (2022, background): “However, during the inference generation, the model generates the next token conditioned on the historical sequences previously generated by itself [69].”
• On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes (2023, background): “While He et al. (2019) argue that this mismatch may not be critical, several papers demonstrate that exposure bias leads to poor text generation (Zhang et al., 2019; Chiang & Chen, 2021; Arora et al., 2022).”