@inproceedings{xiao-etal-2026-lunar-r1,
title = "{L}unar-{R}1: Efficient Reasoning of Large Language Models for Lunar Exploration via Reinforcement Learning",
author = "Xiao, Xin-Yu and
Tian, Ye and
He, Zhixian and
Chen, Qilin and
Wang, Shiqi and
Xia, Qianchen",
editor = "Atanasova, Pepa and
Haffari, Gholamreza and
Wang, Minghan",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-industry.5/",
pages = "50--69",
ISBN = "979-8-89176-460-6",
abstract = "We present Lunar-R1, an 8B-parameter reasoning model tailored for lunar science and mission operations, developed through continued pre-training, supervised fine-tuning, and reinforcement learning. To optimize the accuracy-efficiency trade-off, we introduce Latent Difficulty Perception (LDP), a lightweight mechanism that leverages final-layer hidden states to estimate query difficulty and regulates Chain-of-Thought generation length via difficulty-aware reward modeling. We construct Lunar-QA, an expert-verified benchmark derived from authentic lunar mission materials, and further evaluate on domain-specific tasks spanning astronomy and astrodynamics. Experimental results demonstrate that LDP reduces average token consumption by 38.9{\%} while improving reasoning accuracy. Profiling on an RK3588 NPU yields 0.54 J/token energy efficiency and 380 ms time-to-first-token latency, validating feasibility for lunar surface edge deployment."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xiao-etal-2026-lunar-r1">
<titleInfo>
<title>Lunar-R1: Efficient Reasoning of Large Language Models for Lunar Exploration via Reinforcement Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xin-Yu</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ye</namePart>
<namePart type="family">Tian</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhixian</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qilin</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shiqi</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qianchen</namePart>
<namePart type="family">Xia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Pepa</namePart>
<namePart type="family">Atanasova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gholamreza</namePart>
<namePart type="family">Haffari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Minghan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-460-6</identifier>
</relatedItem>
<abstract>We present Lunar-R1, an 8B-parameter reasoning model tailored for lunar science and mission operations, developed through continued pre-training, supervised fine-tuning, and reinforcement learning. To optimize the accuracy-efficiency trade-off, we introduce Latent Difficulty Perception (LDP), a lightweight mechanism that leverages final-layer hidden states to estimate query difficulty and regulates Chain-of-Thought generation length via difficulty-aware reward modeling. We construct Lunar-QA, an expert-verified benchmark derived from authentic lunar mission materials, and further evaluate on domain-specific tasks spanning astronomy and astrodynamics. Experimental results demonstrate that LDP reduces average token consumption by 38.9% while improving reasoning accuracy. Profiling on an RK3588 NPU yields 0.54 J/token energy efficiency and 380 ms time-to-first-token latency, validating feasibility for lunar surface edge deployment.</abstract>
<identifier type="citekey">xiao-etal-2026-lunar-r1</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-industry.5/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>50</start>
<end>69</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Lunar-R1: Efficient Reasoning of Large Language Models for Lunar Exploration via Reinforcement Learning
%A Xiao, Xin-Yu
%A Tian, Ye
%A He, Zhixian
%A Chen, Qilin
%A Wang, Shiqi
%A Xia, Qianchen
%Y Atanasova, Pepa
%Y Haffari, Gholamreza
%Y Wang, Minghan
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-460-6
%F xiao-etal-2026-lunar-r1
%X We present Lunar-R1, an 8B-parameter reasoning model tailored for lunar science and mission operations, developed through continued pre-training, supervised fine-tuning, and reinforcement learning. To optimize the accuracy-efficiency trade-off, we introduce Latent Difficulty Perception (LDP), a lightweight mechanism that leverages final-layer hidden states to estimate query difficulty and regulates Chain-of-Thought generation length via difficulty-aware reward modeling. We construct Lunar-QA, an expert-verified benchmark derived from authentic lunar mission materials, and further evaluate on domain-specific tasks spanning astronomy and astrodynamics. Experimental results demonstrate that LDP reduces average token consumption by 38.9% while improving reasoning accuracy. Profiling on an RK3588 NPU yields 0.54 J/token energy efficiency and 380 ms time-to-first-token latency, validating feasibility for lunar surface edge deployment.
%U https://aclanthology.org/2026.emnlp-industry.5/
%P 50-69
Markdown (Informal)
[Lunar-R1: Efficient Reasoning of Large Language Models for Lunar Exploration via Reinforcement Learning](https://aclanthology.org/2026.emnlp-industry.5/) (Xiao et al., EMNLP 2026)
ACL