arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.16784cs.LGcs.CLstat.ME
收藏

语言模型中分布偏移下的遗漏变量偏差

Omitted Variable Bias in Language Models Under Distribution Shift

Carnegie Mellon University(卡内基梅隆大学)

Victoria Lin, Louis-Philippe Morency, Eli Ben-Michael

AI总结:

本文研究了语言模型在分布偏移下的遗漏变量偏差问题,提出框架用于评估和优化模型在分布偏移下的性能。

摘要:

尽管在各种任务上表现优异,现代语言模型仍易受分布偏移影响,当评估数据分布与训练数据不同时,表现出脆弱性。本文描述了语言模型中的分布偏移可以分为可观测和不可观测的组成部分,并讨论了现有方法仅处理前者。重要的是,我们发现不可观测变量导致的遗漏变量偏差会损害语言模型的评估和优化。为解决这一挑战,我们引入了一个框架,将遗漏变量的强度映射到分布偏移下语言模型最坏情况泛化性能的界限。在实证实验中,我们显示使用这些界限直接进行语言模型评估和优化,提供了更系统的分布外性能度量,相对于标准分布偏移调整方法,进一步提高了真实的分布外性能,并且当目标分布标签可用时,能够推断出遗漏变量的强度。

英文摘要:

Despite their impressive performance on a wide variety of tasks, modern language models remain susceptible to distribution shifts, exhibiting brittle behavior when evaluated on data that differs in distribution from their training data. In this paper, we describe how distribution shifts in language models can be separated into observable and unobservable components, and we discuss how established approaches for dealing with distribution shift address only the former. Importantly, we identify that the resulting omitted variable bias from unobserved variables can compromise both evaluation and optimization in language models. To address this challenge, we introduce a framework that maps the strength of the omitted variables to bounds on the worst-case generalization performance of language models under distribution shift. In empirical experiments, we show that using these bounds directly in language model evaluation and optimization provides more principled measures of out-of-distribution performance, improves true out-of-distribution performance relative to standard distribution shift adjustment methods, and further enables inference about the strength of the omitted variables when target distribution labels are available.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑