Consistency of a Recurrent Language Model with Respect to Incomplete Decoding

By Sean Welleck, Ilia Kulikov, Jaedeok Kim, Richard Yuanzhe Pang, Kyunghyun Cho. 2020

76 citations recorded by the source corpus.

Despite strong performance on a variety of tasks, neural sequence models trained with maximum likelihood have been shown to exhibit issues such as length bias and degenerate repetition. We study the related issue of receiving infinite-length sequences from a recurrent language model when using common decoding algorithms. To analyze this issue, we first define inconsistency of a decoding algorithm, meaning that the algorithm can yield an infinite-length sequence that has zero probability under the model. We prove that commonly used incomplete decoding algorithms - greedy search, beam search, top-k sampling, and nucleus sampling - are inconsistent, despite the fact that recurrent language models are trained to produce sequences of finite length. Based on these insights, we propose two remedies which address inconsistency: consistent variants of top-k and nucleus sampling, and a self-terminating recurrent language model. Empirical results show that inconsistency occurs in practice, and that the proposed methods prevent inconsistency.

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

In Computer Vision and Pattern Recognition

In Natural Language Processing

Citations in the corpus

Self-Consistency Improves Chain of Thought Reasoning in Language Models (2022, background): “Welleck et al. (2020) use “consistency” to refer to generating an inﬁnite-length sequence in recurrent language models.”
AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts (2021, background): “Exposure bias can also cause LLMs to produce redundant content, in some severe cases repeating the same phrase over and over again [32, 72].”
Maieutic Prompting: Logically Consistent Reasoning with Recursive Explanations (2022, background): “This is in part due to the broader limitations of generative LMs, which assign high probability to unlikely sentences (Welleck et al., 2020; Holtzman et al., 2021) and are sensitive to semantic perturbations (Elazar et al., 2021).”
Text Generation by Learning from Demonstrations (2020, unknown): “10 Qualitatively, we observe frequent degenerations (Holtzman et al., 2020; Welleck et al., 2020a) including repetitions and hallucinations within a sentence generated by MLE-trained model, as shown in Table 5.”
MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontiers (2021, methodology): “However, quantifying these aspects in a principled yet computationally tractable manner is challenging, as the text distributions are high-dimensional and discrete, accessed only through samples or expensive model evaluations [26, 58, 62].”