On Subjective Uncertainty Quantification and Calibration in Natural Language Generation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Ziyu, Holmes, Chris
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917807618260992
author Wang, Ziyu
Holmes, Chris
author_facet Wang, Ziyu
Holmes, Chris
contents Applications of large language models often involve the generation of free-form responses, in which case uncertainty quantification becomes challenging. This is due to the need to identify task-specific uncertainties (e.g., about the semantics) which appears difficult to define in general cases. This work addresses these challenges from a perspective of Bayesian decision theory, starting from the assumption that our utility is characterized by a similarity measure that compares a generated response with a hypothetical true response. We discuss how this assumption enables principled quantification of the model's subjective uncertainty and its calibration. We further derive a measure for epistemic uncertainty, based on a missing data perspective and its characterization as an excess risk. The proposed methods can be applied to black-box language models. We illustrate the methods on question answering and machine translation tasks. Our experiments provide a principled evaluation of task-specific calibration, and demonstrate that epistemic uncertainty offers a promising deferral strategy for efficient data acquisition in in-context learning.
format Preprint
id arxiv_https___arxiv_org_abs_2406_05213
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle On Subjective Uncertainty Quantification and Calibration in Natural Language Generation
Wang, Ziyu
Holmes, Chris
Computation and Language
Artificial Intelligence
Machine Learning
Applications of large language models often involve the generation of free-form responses, in which case uncertainty quantification becomes challenging. This is due to the need to identify task-specific uncertainties (e.g., about the semantics) which appears difficult to define in general cases. This work addresses these challenges from a perspective of Bayesian decision theory, starting from the assumption that our utility is characterized by a similarity measure that compares a generated response with a hypothetical true response. We discuss how this assumption enables principled quantification of the model's subjective uncertainty and its calibration. We further derive a measure for epistemic uncertainty, based on a missing data perspective and its characterization as an excess risk. The proposed methods can be applied to black-box language models. We illustrate the methods on question answering and machine translation tasks. Our experiments provide a principled evaluation of task-specific calibration, and demonstrate that epistemic uncertainty offers a promising deferral strategy for efficient data acquisition in in-context learning.
title On Subjective Uncertainty Quantification and Calibration in Natural Language Generation
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2406.05213