@inproceedings{lv-etal-2025-raise,
title = "{RAISE}: Reinforced Adaptive Instruction Selection For Large Language Models",
author = "Lv, Qingsong and
Li, Yangning and
Lan, Zihua and
Xu, Zishan and
Tang, Jiwei and
Lu, Tingwei and
Li, Yinghui and
Jiang, Wenhao and
Kim, Hong-Gee and
Zheng, Hai-Tao and
Yu, Philip S.",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-emnlp.628/",
pages = "11708--11723",
ISBN = "979-8-89176-335-7",
abstract = "Instruction tuning of large language models (LLMs) benefits more from a handful of high-quality examples than from hordes of low-quality ones. Existing selection methods typically rely on static, heuristic quality scores and are executed only once before training. Consequently, they neither adapt to the changing state of the model nor target downstream objectives, leaving substantial room for optimization. We propose RAISE (**R**einforced **A**daptive **I**nstruction **SE**lection), a *dynamic*, *task-driven* framework that integrates selection into every training step. At each step, RAISE estimates the expected contribution of each candidate instruction to task performance and admits only the most helpful. By modeling this process as sequential decision making, we optimize the selector with reinforcement learning, yielding an interpretable policy specialized for the target task. Extensive experiments show that RAISE reaches comparable or better results than full-data training while updating only 1{\%} of the steps, demonstrating both high efficacy and significant computational savings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lv-etal-2025-raise">
<titleInfo>
<title>RAISE: Reinforced Adaptive Instruction Selection For Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Qingsong</namePart>
<namePart type="family">Lv</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yangning</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zihua</namePart>
<namePart type="family">Lan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zishan</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiwei</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tingwei</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yinghui</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenhao</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hong-Gee</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hai-Tao</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philip</namePart>
<namePart type="given">S</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-335-7</identifier>
</relatedItem>
<abstract>Instruction tuning of large language models (LLMs) benefits more from a handful of high-quality examples than from hordes of low-quality ones. Existing selection methods typically rely on static, heuristic quality scores and are executed only once before training. Consequently, they neither adapt to the changing state of the model nor target downstream objectives, leaving substantial room for optimization. We propose RAISE (**R**einforced **A**daptive **I**nstruction **SE**lection), a *dynamic*, *task-driven* framework that integrates selection into every training step. At each step, RAISE estimates the expected contribution of each candidate instruction to task performance and admits only the most helpful. By modeling this process as sequential decision making, we optimize the selector with reinforcement learning, yielding an interpretable policy specialized for the target task. Extensive experiments show that RAISE reaches comparable or better results than full-data training while updating only 1% of the steps, demonstrating both high efficacy and significant computational savings.</abstract>
<identifier type="citekey">lv-etal-2025-raise</identifier>
<location>
<url>https://aclanthology.org/2025.findings-emnlp.628/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>11708</start>
<end>11723</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T RAISE: Reinforced Adaptive Instruction Selection For Large Language Models
%A Lv, Qingsong
%A Li, Yangning
%A Lan, Zihua
%A Xu, Zishan
%A Tang, Jiwei
%A Lu, Tingwei
%A Li, Yinghui
%A Jiang, Wenhao
%A Kim, Hong-Gee
%A Zheng, Hai-Tao
%A Yu, Philip S.
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F lv-etal-2025-raise
%X Instruction tuning of large language models (LLMs) benefits more from a handful of high-quality examples than from hordes of low-quality ones. Existing selection methods typically rely on static, heuristic quality scores and are executed only once before training. Consequently, they neither adapt to the changing state of the model nor target downstream objectives, leaving substantial room for optimization. We propose RAISE (**R**einforced **A**daptive **I**nstruction **SE**lection), a *dynamic*, *task-driven* framework that integrates selection into every training step. At each step, RAISE estimates the expected contribution of each candidate instruction to task performance and admits only the most helpful. By modeling this process as sequential decision making, we optimize the selector with reinforcement learning, yielding an interpretable policy specialized for the target task. Extensive experiments show that RAISE reaches comparable or better results than full-data training while updating only 1% of the steps, demonstrating both high efficacy and significant computational savings.
%U https://aclanthology.org/2025.findings-emnlp.628/
%P 11708-11723
Markdown (Informal)
[RAISE: Reinforced Adaptive Instruction Selection For Large Language Models](https://aclanthology.org/2025.findings-emnlp.628/) (Lv et al., Findings 2025)
ACL
- Qingsong Lv, Yangning Li, Zihua Lan, Zishan Xu, Jiwei Tang, Tingwei Lu, Yinghui Li, Wenhao Jiang, Hong-Gee Kim, Hai-Tao Zheng, and Philip S. Yu. 2025. RAISE: Reinforced Adaptive Instruction Selection For Large Language Models. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 11708–11723, Suzhou, China. Association for Computational Linguistics.