Differentiable Instruction Optimization for Cross-Task Generalization

Autor:	Isonuma, Masaru, Mori, Junichiro, Sakata, Ichiro
Rok vydání:	2023
Předmět:	Computer Science - Computation and Language Computer Science - Artificial Intelligence Computer Science - Machine Learning
Druh dokumentu:	Working Paper
Popis:	Instruction tuning has been attracting much attention to achieve generalization ability across a wide variety of tasks. Although various types of instructions have been manually created for instruction tuning, it is still unclear what kind of instruction is optimal to obtain cross-task generalization ability. This work presents instruction optimization, which optimizes training instructions with respect to generalization ability. Rather than manually tuning instructions, we introduce learnable instructions and optimize them with gradient descent by leveraging bilevel optimization. Experimental results show that the learned instruction enhances the diversity of instructions and improves the generalization ability compared to using only manually created instructions. Comment: 14pages, 6 figures, accepted for Findings of ACL2023
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2306.10098 Zobrazit plný text záznamu View this record from Arxiv