Next Article in Journal
A Linguistic q-Rung Orthopair ELECTRE II Algorithm for Fuzzy Multi-Criteria Ontology Ranking
Previous Article in Journal
LizAI XT—AI-Accelerated Management Platform for Complex Healthcare Data at Scale, Beyond EMR/EHR and Dashboards
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
BDCC
Volume 9
Issue 11
10.3390/bdcc9110276
BDCC-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
This is an early access version, the complete PDF, HTML, and XML versions will be available soon.
Article

Demo-ToT: Enhancing the Reasoning Capabilities of AI Agent via Improved Demonstrations Retrieval Strategy

by
Jiahui Li
,
Bangbang Ren
,
Mengmeng Zhang
* and
Honghui Chen
National Key Laboratory of Information Systems Engineering, National University of Defense Technology, Changsha 410073, China
*
Author to whom correspondence should be addressed.
Big Data Cogn. Comput. 2025, 9(11), 276; https://doi.org/10.3390/bdcc9110276
Submission received: 1 September 2025 / Revised: 16 October 2025 / Accepted: 31 October 2025 / Published: 2 November 2025
(This article belongs to the Special Issue Artificial Intelligence (AI) and Natural Language Processing (NLP))
Versions Notes

Abstract

Innovative reasoning frameworks have been proposed to enhance the reasoning capabilities of AI agents, improving their performance in various tasks. However, most existing research has focused on enhancing designing frameworks for LLMs, with limited attention on leveraging in-context learning to boost their reasoning power. This paper proposes a novel approach, Demo-ToT, which enhances the Tree-of-Thought (ToT) reasoning framework by dynamically retrieving relevant demonstrations to improve reasoning accuracy. Various demonstration retrieval strategies, including vector similarity, sparse retrieval, and string similarity, were explored to identify the most effective methods for optimizing LLM performance. Experiments conducted across multiple benchmarks and language models of varying sizes demonstrated that Demo-ToT substantially enhanced the reasoning ability of smaller LLMs, achieving performance comparable to or even surpassing that of much larger models such as GPT-4.
Keywords: large language models; demonstration learning; tree-of-thought large language models; demonstration learning; tree-of-thought

Share and Cite

MDPI and ACS Style

Li, J.; Ren, B.; Zhang, M.; Chen, H. Demo-ToT: Enhancing the Reasoning Capabilities of AI Agent via Improved Demonstrations Retrieval Strategy. Big Data Cogn. Comput. 2025, 9, 276. https://doi.org/10.3390/bdcc9110276

AMA Style

Li J, Ren B, Zhang M, Chen H. Demo-ToT: Enhancing the Reasoning Capabilities of AI Agent via Improved Demonstrations Retrieval Strategy. Big Data and Cognitive Computing. 2025; 9(11):276. https://doi.org/10.3390/bdcc9110276

Chicago/Turabian Style

Li, Jiahui, Bangbang Ren, Mengmeng Zhang, and Honghui Chen. 2025. "Demo-ToT: Enhancing the Reasoning Capabilities of AI Agent via Improved Demonstrations Retrieval Strategy" Big Data and Cognitive Computing 9, no. 11: 276. https://doi.org/10.3390/bdcc9110276

APA Style

Li, J., Ren, B., Zhang, M., & Chen, H. (2025). Demo-ToT: Enhancing the Reasoning Capabilities of AI Agent via Improved Demonstrations Retrieval Strategy. Big Data and Cognitive Computing, 9(11), 276. https://doi.org/10.3390/bdcc9110276

Article Metrics

Back to TopTop