利用跨任务最近邻进行数据高效微调
Data-Efficient Finetuning Using Cross-Task Nearest Neighbors
AI总结：
本文提出利用少量未标注目标样本从多任务提示数据池中检索跨任务最近邻进行微调，以极低数据量实现高效任务适配并优于强基线。
摘要：
为感兴趣的任务获取标注数据来训练模型通常代价高昂。先前工作表明，在带有任务描述（提示）的多任务数据上训练模型，能够有效地将知识迁移到新任务。为了高效构建任务特定模型，我们假设可以访问少量（32-1000 个）未标注的目标任务样本，并利用这些样本从带有提示的大规模多任务数据池中检索最相似的已标注样本。与当前在均匀采样的带提示多任务数据（例如 FLAN、T0）上微调模型的做法相比，我们在跨任务最近邻上进行微调的方法在数据效率上显著更高。在没有任何目标任务标注数据的情况下，仅使用 P3 数据池中 2% 的数据，我们的模型就在 14 个代表留出任务（包括法律和科学文档问答）的数据集中的 12 个上，以 3-30% 的优势超过了在所有可用数据上训练的强基线模型。类似地，在来自 SuperNaturalInstructions 的跨任务最近邻（约占该数据池的 5%）上训练的模型，在该数据池的 12 个留出任务上取得了与最先进模型相当的性能。此外，我们的方法所生成的模型还为在目标任务数据上进行少样本微调提供了比单一多任务微调模型更好的初始化，具体表现为在 8 个数据集上相对于少样本微调的 T0-3B 模型有 2-23% 的相对提升。
英文摘要：
Obtaining labeled data to train a model for a task of interest is often expensive. Prior work shows training models on multitask data augmented with task descriptions (prompts) effectively transfers knowledge to new tasks. Towards efficiently building task-specific models, we assume access to a small number (32-1000) of unlabeled target-task examples and use those to retrieve the most similar labeled examples from a large pool of multitask data augmented with prompts. Compared to the current practice of finetuning models on uniformly sampled prompted multitask data (e.g.: FLAN, T0), our approach of finetuning on cross-task nearest neighbors is significantly more data-efficient. Using only 2% of the data from the P3 pool without any labeled target-task data, our models outperform strong baselines trained on all available data by 3-30% on 12 out of 14 datasets representing held-out tasks including legal and scientific document QA. Similarly, models trained on cross-task nearest neighbors from SuperNaturalInstructions, representing about 5% of the pool, obtain comparable performance to state-of-the-art models on 12 held-out tasks from that pool. Moreover, the models produced by our approach also provide a better initialization than single multitask finetuned models for few-shot finetuning on target-task data, as shown by a 2-23% relative improvement over few-shot finetuned T0-3B models on 8 datasets.
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。
