@inproceedings{zhu-etal-2025-divide,
title = "Divide-Then-Aggregate: An Efficient Tool Learning Method via Parallel Tool Invocation",
author = "Zhu, Dongsheng and
Shi, Weixian and
Shi, Zhengliang and
Ren, Zhaochun and
Wang, Shuaiqiang and
Yan, Lingyong and
Yin, Dawei",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.1401/",
doi = "10.18653/v1/2025.acl-long.1401",
pages = "28859--28875",
ISBN = "979-8-89176-251-0",
abstract = "While Large Language Models (LLMs) demonstrate remarkable capabilities, their ability to autonomously execute complex real-world tasks remains limited. Accordingly, tool learning has emerged to enable LLMs to effectively leverage external tools to extend their capabilities. Current tool-learning paradigms like CoT/ReAct employ sequential tool invocation but suffer from constrained perception and inadequate task planning. Alternative approaches using search-based decision trees incur substantial computational overhead. To address these limitations, we propose DTA-Llama (Divide-Then-Aggregate Llama), a novel parallel tool invocation framework featuring: (1) A Directed Acyclic Graph (DAG) structure that transformed from traditional tree-based tool search paths, enabling parallel execution and contributing high-quality training data; (2) A process-thread-inspired inference mechanism that iteratively decomposes tasks into parallel tool-using subtasks while aggregating results for subsequent decisions. Experimental results show that our approach substantially enhances task performance while reducing token consumption and inference time. Llama2-7B, using our method, is comparable to the official parallel function calling method of GPT-3.5. The relevant code, dataset, and model weights are available at https://corn0205.github.io/."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhu-etal-2025-divide">
<titleInfo>
<title>Divide-Then-Aggregate: An Efficient Tool Learning Method via Parallel Tool Invocation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dongsheng</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weixian</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengliang</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhaochun</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuaiqiang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lingyong</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dawei</namePart>
<namePart type="family">Yin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-251-0</identifier>
</relatedItem>
<abstract>While Large Language Models (LLMs) demonstrate remarkable capabilities, their ability to autonomously execute complex real-world tasks remains limited. Accordingly, tool learning has emerged to enable LLMs to effectively leverage external tools to extend their capabilities. Current tool-learning paradigms like CoT/ReAct employ sequential tool invocation but suffer from constrained perception and inadequate task planning. Alternative approaches using search-based decision trees incur substantial computational overhead. To address these limitations, we propose DTA-Llama (Divide-Then-Aggregate Llama), a novel parallel tool invocation framework featuring: (1) A Directed Acyclic Graph (DAG) structure that transformed from traditional tree-based tool search paths, enabling parallel execution and contributing high-quality training data; (2) A process-thread-inspired inference mechanism that iteratively decomposes tasks into parallel tool-using subtasks while aggregating results for subsequent decisions. Experimental results show that our approach substantially enhances task performance while reducing token consumption and inference time. Llama2-7B, using our method, is comparable to the official parallel function calling method of GPT-3.5. The relevant code, dataset, and model weights are available at https://corn0205.github.io/.</abstract>
<identifier type="citekey">zhu-etal-2025-divide</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.1401</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.1401/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>28859</start>
<end>28875</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Divide-Then-Aggregate: An Efficient Tool Learning Method via Parallel Tool Invocation
%A Zhu, Dongsheng
%A Shi, Weixian
%A Shi, Zhengliang
%A Ren, Zhaochun
%A Wang, Shuaiqiang
%A Yan, Lingyong
%A Yin, Dawei
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F zhu-etal-2025-divide
%X While Large Language Models (LLMs) demonstrate remarkable capabilities, their ability to autonomously execute complex real-world tasks remains limited. Accordingly, tool learning has emerged to enable LLMs to effectively leverage external tools to extend their capabilities. Current tool-learning paradigms like CoT/ReAct employ sequential tool invocation but suffer from constrained perception and inadequate task planning. Alternative approaches using search-based decision trees incur substantial computational overhead. To address these limitations, we propose DTA-Llama (Divide-Then-Aggregate Llama), a novel parallel tool invocation framework featuring: (1) A Directed Acyclic Graph (DAG) structure that transformed from traditional tree-based tool search paths, enabling parallel execution and contributing high-quality training data; (2) A process-thread-inspired inference mechanism that iteratively decomposes tasks into parallel tool-using subtasks while aggregating results for subsequent decisions. Experimental results show that our approach substantially enhances task performance while reducing token consumption and inference time. Llama2-7B, using our method, is comparable to the official parallel function calling method of GPT-3.5. The relevant code, dataset, and model weights are available at https://corn0205.github.io/.
%R 10.18653/v1/2025.acl-long.1401
%U https://aclanthology.org/2025.acl-long.1401/
%U https://doi.org/10.18653/v1/2025.acl-long.1401
%P 28859-28875
Markdown (Informal)
[Divide-Then-Aggregate: An Efficient Tool Learning Method via Parallel Tool Invocation](https://aclanthology.org/2025.acl-long.1401/) (Zhu et al., ACL 2025)
ACL