让语言模型生成更多合理且多样的句子延续。
Multiple Choice Learning of Low-Rank Adapters for Language Modeling
- 用低秩适配+多选学习，让模型在推理时考虑多种可能的接续。
- 在图文描述和机器翻译任务中，生成结果既多样又准确。
- 适合需要丰富输出的场景，如对话生成、内容创作。
我们提出LoRA-MCL，一种扩展语言模型下一词预测的训练方法，旨在推理时生成多样且合理的句子延续。传统语言建模本质上是病态问题：给定上下文后，多个未来可能同样合理。该方法结合多选学习（MCL）与胜者为王损失，通过低秩适配高效处理歧义。我们从理论上解释了将MCL应用于语言建模的合理性，假设数据来自混合分布。通过马尔可夫链混合示例进行说明，并在音频描述、视觉描述及机器翻译任务上验证，结果显示生成结果兼具高多样性与相关性。代码已开源，适用于多种语言模型。
原文摘要 · Abstract (English)
We propose LoRA-MCL, a training scheme that extends next-token prediction in language models with a method designed to decode diverse, plausible sentence continuations at inference time. Traditional language modeling is an intrinsically ill-posed problem: given a context, multiple futures may be equally plausible. Our approach leverages Multiple Choice Learning (MCL) and the winner-takes-all loss to efficiently handle ambiguity through Low-Rank Adaptation. We provide a theoretical interpretation of applying MCL to language modeling, assuming the data is generated from a mixture of distributions. We illustrate the proposed approach using mixtures of Markov chains. We then demonstrate with experiments on audio and visual captioning, as well as machine translation, that our method achieves high diversity and relevance in generated outputs. We release the code for applying LoRA-MCL to a wide range of language models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。