@inproceedings{xing-etal-2026-compute,
title = "Compute Allocation in Evolutionary Search: From Depth{--}Breadth to Multi-Armed Bandits",
author = "Xing, Sixue and
He, Haoyu and
Wu, Kerui and
Yang, Zhuo and
Luo, Haozheng and
Fu, Tianfan and
Nagarajan, Aarthy",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1056/",
pages = "24224--24251",
ISBN = "979-8-89176-459-0",
abstract = "LLM-guided evolutionary search (Evolve systems) has reached state-of-the-art results on mathematical and combinatorial tasks, yet most existing systems report only the best of many runs and leave the run-to-run distribution undocumented. We ask how a fixed budget of LLM calls should be allocated, and how reliably a single run reaches the reported numbers. Sweeping the depth{--}breadth grid over five models and three tasks, we identify two empirical regularities: a fitness{--}compute envelope along which capability ordering largely collapses on effective FLOPs, and a bilinear depth{--}breadth fit with task-specific interaction; both are gated by model{--}task capability. Motivated by these regularities, we propose BaSE (Bandit-based Self-Evolving), a multi-armed bandit that allocates LLM calls across parallel trajectories. Without changing the model, prompt, or evaluator, BaSE improves mean fitness by 12.3{\%} over the strongest island-protocol baseline across 8 (model, task) cells, with the largest gains on high-variance settings: a reliability gain from allocation alone."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xing-etal-2026-compute">
<titleInfo>
<title>Compute Allocation in Evolutionary Search: From Depth–Breadth to Multi-Armed Bandits</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sixue</namePart>
<namePart type="family">Xing</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoyu</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kerui</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhuo</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haozheng</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianfan</namePart>
<namePart type="family">Fu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aarthy</namePart>
<namePart type="family">Nagarajan</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</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</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">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>LLM-guided evolutionary search (Evolve systems) has reached state-of-the-art results on mathematical and combinatorial tasks, yet most existing systems report only the best of many runs and leave the run-to-run distribution undocumented. We ask how a fixed budget of LLM calls should be allocated, and how reliably a single run reaches the reported numbers. Sweeping the depth–breadth grid over five models and three tasks, we identify two empirical regularities: a fitness–compute envelope along which capability ordering largely collapses on effective FLOPs, and a bilinear depth–breadth fit with task-specific interaction; both are gated by model–task capability. Motivated by these regularities, we propose BaSE (Bandit-based Self-Evolving), a multi-armed bandit that allocates LLM calls across parallel trajectories. Without changing the model, prompt, or evaluator, BaSE improves mean fitness by 12.3% over the strongest island-protocol baseline across 8 (model, task) cells, with the largest gains on high-variance settings: a reliability gain from allocation alone.</abstract>
<identifier type="citekey">xing-etal-2026-compute</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1056/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>24224</start>
<end>24251</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Compute Allocation in Evolutionary Search: From Depth–Breadth to Multi-Armed Bandits
%A Xing, Sixue
%A He, Haoyu
%A Wu, Kerui
%A Yang, Zhuo
%A Luo, Haozheng
%A Fu, Tianfan
%A Nagarajan, Aarthy
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F xing-etal-2026-compute
%X LLM-guided evolutionary search (Evolve systems) has reached state-of-the-art results on mathematical and combinatorial tasks, yet most existing systems report only the best of many runs and leave the run-to-run distribution undocumented. We ask how a fixed budget of LLM calls should be allocated, and how reliably a single run reaches the reported numbers. Sweeping the depth–breadth grid over five models and three tasks, we identify two empirical regularities: a fitness–compute envelope along which capability ordering largely collapses on effective FLOPs, and a bilinear depth–breadth fit with task-specific interaction; both are gated by model–task capability. Motivated by these regularities, we propose BaSE (Bandit-based Self-Evolving), a multi-armed bandit that allocates LLM calls across parallel trajectories. Without changing the model, prompt, or evaluator, BaSE improves mean fitness by 12.3% over the strongest island-protocol baseline across 8 (model, task) cells, with the largest gains on high-variance settings: a reliability gain from allocation alone.
%U https://aclanthology.org/2026.emnlp-main.1056/
%P 24224-24251
Markdown (Informal)
[Compute Allocation in Evolutionary Search: From Depth–Breadth to Multi-Armed Bandits](https://aclanthology.org/2026.emnlp-main.1056/) (Xing et al., EMNLP 2026)
ACL