CYCLE-INSTRUCT：通过双重自训练与循环一致性实现完全无种子指令微调
CYCLE-INSTRUCT: Fully Seed-Free Instruction Tuning via Dual Self-Training and Cycle Consistency
AI总结：
本文提出Cycle-Instruct框架，通过答案生成器与问题生成器的双重自训练和循环一致性机制，仅利用原始无标注文本实现完全无种子的指令微调，在多个数据轨道上取得优于种子驱动回译基线且媲美强监督方法的性能。
摘要：
指令微调对于使大语言模型（LLMs）与人类意图对齐至关重要，但当前方法通常依赖昂贵的人工标注种子数据或强大的外部教师模型。虽然指令回译技术降低了这种依赖，但它们从根本上仍受限于初始种子集，这限制了完全自动化，引入了偏差，并可能导致对无标注语料的低效利用。本文提出Cycle-Instruct，一个实现完全无种子指令微调的新框架。受循环一致性启发，Cycle-Instruct采用双重自训练循环，其中两个模型——答案生成器和问题生成器——仅从原始无标注文本中引导训练。这两个模型通过从对方生成的伪标签中重建原始文本片段来相互监督，从而在没有任何人工提供种子的情况下有效地从数据的内在结构中学习。我们在四个不同的数据轨道上展示了Cycle-Instruct的有效性，包括通用指令遵循、领域特定任务、对话日志和纯文本。大量实验表明，Cycle-Instruct不仅优于依赖种子的回译基线方法，而且达到了与强监督方法相当的性能。
英文摘要：
Instruction tuning is vital for aligning large language models (LLMs) with human intent, but current methods typically rely on costly human-annotated seed data or powerful external teacher models. While instruction back-translation techniques reduce this dependency, they remain fundamentally tethered to an initial seed set, which limits full automation, introduces biases, and can lead to inefficient use of unlabeled corpora. In this paper, we propose Cycle-Instruct, a novel framework that achieves fully seed-free instruction tuning. Inspired by cycle consistency, Cycle-Instruct employs a dual self-training loop where two models-an answer generator and a question generator-are bootstrapped solely from raw, unlabeled text. These models mutually supervise each other by reconstructing original text segments from their counterpart's generated pseudo-labels, effectively learning from the intrinsic structure of the data without any human-provided seeds. We demonstrate Cycle-Instruct's efficacy across four diverse data tracks, including general instruction-following, domain-specific tasks, dialogue logs, and plain text. Our extensive experiments show that Cycle-Instruct not only outperforms seed-driven back-translation baselines but also achieves performance comparable to strongly supervised methods.
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。
