@inproceedings{yoshida-etal-2025-batch,
title = "Batch-wise Convergent Pre-training: Step-by-Step Learning Inspired by Child Language Development",
author = "Yoshida, Ko and
Shiono, Daiki and
Sato, Kai and
Miura, Toko and
Furuhashi, Momoka and
Suzuki, Jun",
editor = "Charpentier, Lucas and
Choshen, Leshem and
Cotterell, Ryan and
Gul, Mustafa Omer and
Hu, Michael Y. and
Liu, Jing and
Jumelet, Jaap and
Linzen, Tal and
Mueller, Aaron and
Ross, Candace and
Shah, Raj Sanjay and
Warstadt, Alex and
Wilcox, Ethan Gotlieb and
Williams, Adina",
booktitle = "Proceedings of the First BabyLM Workshop",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.babylm-main.36/",
pages = "508--524",
ISBN = "TODO",
abstract = "Human children acquire language from a substantially smaller amount of linguistic input than that typically required for training large language models (LLMs). This gap motivates the search for more efficient pre-training methods. Inspired by child development, curriculum learning, which progresses from simple to complex data, has been widely adopted. In this study, we propose a pre-training framework that mirrors child language acquisition, advancing step by step from words to sentences while retaining prior knowledge. We investigate whether this improves retention and efficiency under limited resources. Our approach is implemented through four components: (i) a curriculum-aligned dataset, (ii) a batch-wise convergence loop, (iii) a distance-controlled loss to mitigate forgetting, and (iv) a constraint-controlled optimizer for stability. Experiments on the BabyLM benchmark show that the proposed method performs slightly below the official baselines in overall accuracy, with larger gaps on grammar-oriented evaluations such as BLiMP. Nonetheless, it yields small but consistent gains on morphology- and discourse-related tasks (e.g., WUG-ADJ, Entity Tracking), suggesting that the approach affects different linguistic aspects unevenly under limited data conditions."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yoshida-etal-2025-batch">
<titleInfo>
<title>Batch-wise Convergent Pre-training: Step-by-Step Learning Inspired by Child Language Development</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ko</namePart>
<namePart type="family">Yoshida</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daiki</namePart>
<namePart type="family">Shiono</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai</namePart>
<namePart type="family">Sato</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Toko</namePart>
<namePart type="family">Miura</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Momoka</namePart>
<namePart type="family">Furuhashi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</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 First BabyLM Workshop</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lucas</namePart>
<namePart type="family">Charpentier</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leshem</namePart>
<namePart type="family">Choshen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ryan</namePart>
<namePart type="family">Cotterell</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mustafa</namePart>
<namePart type="given">Omer</namePart>
<namePart type="family">Gul</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Michael</namePart>
<namePart type="given">Y</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaap</namePart>
<namePart type="family">Jumelet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tal</namePart>
<namePart type="family">Linzen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aaron</namePart>
<namePart type="family">Mueller</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Candace</namePart>
<namePart type="family">Ross</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raj</namePart>
<namePart type="given">Sanjay</namePart>
<namePart type="family">Shah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alex</namePart>
<namePart type="family">Warstadt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ethan</namePart>
<namePart type="given">Gotlieb</namePart>
<namePart type="family">Wilcox</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Adina</namePart>
<namePart type="family">Williams</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">TODO</identifier>
</relatedItem>
<abstract>Human children acquire language from a substantially smaller amount of linguistic input than that typically required for training large language models (LLMs). This gap motivates the search for more efficient pre-training methods. Inspired by child development, curriculum learning, which progresses from simple to complex data, has been widely adopted. In this study, we propose a pre-training framework that mirrors child language acquisition, advancing step by step from words to sentences while retaining prior knowledge. We investigate whether this improves retention and efficiency under limited resources. Our approach is implemented through four components: (i) a curriculum-aligned dataset, (ii) a batch-wise convergence loop, (iii) a distance-controlled loss to mitigate forgetting, and (iv) a constraint-controlled optimizer for stability. Experiments on the BabyLM benchmark show that the proposed method performs slightly below the official baselines in overall accuracy, with larger gaps on grammar-oriented evaluations such as BLiMP. Nonetheless, it yields small but consistent gains on morphology- and discourse-related tasks (e.g., WUG-ADJ, Entity Tracking), suggesting that the approach affects different linguistic aspects unevenly under limited data conditions.</abstract>
<identifier type="citekey">yoshida-etal-2025-batch</identifier>
<location>
<url>https://aclanthology.org/2025.babylm-main.36/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>508</start>
<end>524</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Batch-wise Convergent Pre-training: Step-by-Step Learning Inspired by Child Language Development
%A Yoshida, Ko
%A Shiono, Daiki
%A Sato, Kai
%A Miura, Toko
%A Furuhashi, Momoka
%A Suzuki, Jun
%Y Charpentier, Lucas
%Y Choshen, Leshem
%Y Cotterell, Ryan
%Y Gul, Mustafa Omer
%Y Hu, Michael Y.
%Y Liu, Jing
%Y Jumelet, Jaap
%Y Linzen, Tal
%Y Mueller, Aaron
%Y Ross, Candace
%Y Shah, Raj Sanjay
%Y Warstadt, Alex
%Y Wilcox, Ethan Gotlieb
%Y Williams, Adina
%S Proceedings of the First BabyLM Workshop
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ TODO
%F yoshida-etal-2025-batch
%X Human children acquire language from a substantially smaller amount of linguistic input than that typically required for training large language models (LLMs). This gap motivates the search for more efficient pre-training methods. Inspired by child development, curriculum learning, which progresses from simple to complex data, has been widely adopted. In this study, we propose a pre-training framework that mirrors child language acquisition, advancing step by step from words to sentences while retaining prior knowledge. We investigate whether this improves retention and efficiency under limited resources. Our approach is implemented through four components: (i) a curriculum-aligned dataset, (ii) a batch-wise convergence loop, (iii) a distance-controlled loss to mitigate forgetting, and (iv) a constraint-controlled optimizer for stability. Experiments on the BabyLM benchmark show that the proposed method performs slightly below the official baselines in overall accuracy, with larger gaps on grammar-oriented evaluations such as BLiMP. Nonetheless, it yields small but consistent gains on morphology- and discourse-related tasks (e.g., WUG-ADJ, Entity Tracking), suggesting that the approach affects different linguistic aspects unevenly under limited data conditions.
%U https://aclanthology.org/2025.babylm-main.36/
%P 508-524
Markdown (Informal)
[Batch-wise Convergent Pre-training: Step-by-Step Learning Inspired by Child Language Development](https://aclanthology.org/2025.babylm-main.36/) (Yoshida et al., BabyLM 2025)
ACL