@inproceedings{shi-etal-2025-reasoning,
title = "Reasoning under Uncertainty: Efficient {LLM} Inference via Unsupervised Confidence Dilution and Convergent Adaptive Sampling",
author = "Shi, Zhenning and
Zhu, Yijia and
Xie, Yi and
Shi, Junhan and
Xie, Guorui and
Zhang, Haotian and
Jiang, Yong and
Miao, Congcong and
Li, Qing",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.emnlp-main.1638/",
doi = "10.18653/v1/2025.emnlp-main.1638",
pages = "32204--32218",
ISBN = "979-8-89176-332-6",
abstract = "Large language models (LLMs) excel at complex reasoning tasks but often suffer from overconfidence and computational inefficiency due to fixed computation budgets and miscalibrated confidence estimates. We present a novel framework for computationally efficient, trustworthy reasoning under uncertainty, introducing two complementary techniques: Diversity-Aware Self-Signal Dilution (DASD) and Convergent Adaptive Weighted Sampling (CAWS). DASD operates in an unsupervised manner to dilute overconfident, semantically redundant reasoning paths, thereby producing better-calibrated internal confidence estimates. CAWS dynamically allocates computational resources at inference time by aggregating these signals and terminating computation once answer dominance and stability are achieved. Comprehensive experiments across three reasoning datasets demonstrate that our approach maintains accuracy levels while achieving over 70{\%} reduction in inference cost, surpassing competitive baselines. Our framework provides a scalable, unsupervised solution for reliable and efficient LLM reasoning."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shi-etal-2025-reasoning">
<titleInfo>
<title>Reasoning under Uncertainty: Efficient LLM Inference via Unsupervised Confidence Dilution and Convergent Adaptive Sampling</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhenning</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yijia</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junhan</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guorui</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haotian</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yong</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Congcong</namePart>
<namePart type="family">Miao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qing</namePart>
<namePart type="family">Li</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>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-332-6</identifier>
</relatedItem>
<abstract>Large language models (LLMs) excel at complex reasoning tasks but often suffer from overconfidence and computational inefficiency due to fixed computation budgets and miscalibrated confidence estimates. We present a novel framework for computationally efficient, trustworthy reasoning under uncertainty, introducing two complementary techniques: Diversity-Aware Self-Signal Dilution (DASD) and Convergent Adaptive Weighted Sampling (CAWS). DASD operates in an unsupervised manner to dilute overconfident, semantically redundant reasoning paths, thereby producing better-calibrated internal confidence estimates. CAWS dynamically allocates computational resources at inference time by aggregating these signals and terminating computation once answer dominance and stability are achieved. Comprehensive experiments across three reasoning datasets demonstrate that our approach maintains accuracy levels while achieving over 70% reduction in inference cost, surpassing competitive baselines. Our framework provides a scalable, unsupervised solution for reliable and efficient LLM reasoning.</abstract>
<identifier type="citekey">shi-etal-2025-reasoning</identifier>
<identifier type="doi">10.18653/v1/2025.emnlp-main.1638</identifier>
<location>
<url>https://aclanthology.org/2025.emnlp-main.1638/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>32204</start>
<end>32218</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Reasoning under Uncertainty: Efficient LLM Inference via Unsupervised Confidence Dilution and Convergent Adaptive Sampling
%A Shi, Zhenning
%A Zhu, Yijia
%A Xie, Yi
%A Shi, Junhan
%A Xie, Guorui
%A Zhang, Haotian
%A Jiang, Yong
%A Miao, Congcong
%A Li, Qing
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-332-6
%F shi-etal-2025-reasoning
%X Large language models (LLMs) excel at complex reasoning tasks but often suffer from overconfidence and computational inefficiency due to fixed computation budgets and miscalibrated confidence estimates. We present a novel framework for computationally efficient, trustworthy reasoning under uncertainty, introducing two complementary techniques: Diversity-Aware Self-Signal Dilution (DASD) and Convergent Adaptive Weighted Sampling (CAWS). DASD operates in an unsupervised manner to dilute overconfident, semantically redundant reasoning paths, thereby producing better-calibrated internal confidence estimates. CAWS dynamically allocates computational resources at inference time by aggregating these signals and terminating computation once answer dominance and stability are achieved. Comprehensive experiments across three reasoning datasets demonstrate that our approach maintains accuracy levels while achieving over 70% reduction in inference cost, surpassing competitive baselines. Our framework provides a scalable, unsupervised solution for reliable and efficient LLM reasoning.
%R 10.18653/v1/2025.emnlp-main.1638
%U https://aclanthology.org/2025.emnlp-main.1638/
%U https://doi.org/10.18653/v1/2025.emnlp-main.1638
%P 32204-32218
Markdown (Informal)
[Reasoning under Uncertainty: Efficient LLM Inference via Unsupervised Confidence Dilution and Convergent Adaptive Sampling](https://aclanthology.org/2025.emnlp-main.1638/) (Shi et al., EMNLP 2025)
ACL
- Zhenning Shi, Yijia Zhu, Yi Xie, Junhan Shi, Guorui Xie, Haotian Zhang, Yong Jiang, Congcong Miao, and Qing Li. 2025. Reasoning under Uncertainty: Efficient LLM Inference via Unsupervised Confidence Dilution and Convergent Adaptive Sampling. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 32204–32218, Suzhou, China. Association for Computational Linguistics.