<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zheng-etal-2024-breaking">
    <titleInfo>
        <title>Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Wenzhen</namePart>
        <namePart type="family">Zheng</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Wenbo</namePart>
        <namePart type="family">Pan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xu</namePart>
        <namePart type="family">Xu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Libo</namePart>
        <namePart type="family">Qin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Li</namePart>
        <namePart type="family">Yue</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ming</namePart>
        <namePart type="family">Zhou</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2024-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yaser</namePart>
            <namePart type="family">Al-Onaizan</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mohit</namePart>
            <namePart type="family">Bansal</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Yun-Nung</namePart>
            <namePart type="family">Chen</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Miami, Florida, USA</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>In recent years, Large Language Models (LLMs) have made significant strides towards Artificial General Intelligence. However, training these models from scratch requires substantial computational resources and vast amounts of text data. In this paper, we explores an alternative approach to constructing a LLM for a new language by continually pre-training (CPT) from existing pre-trained LLMs, instead of using randomly initialized parameters. Based on parallel experiments on 40 model sizes ranging from 40M to 5B parameters, we find that 1) CPT converges faster and saves significant resources in a scalable manner. 2) CPT adheres to an extended scaling law derived from with a joint data-parameter scaling term. 3) The compute-optimal data-parameter allocation for CPT markedly differs based on our estimated scaling factors. 4) The effectiveness of transfer scale is influenced by training duration and linguistic properties, while robust to data replaying, a method that effectively mitigates catastrophic forgetting in CPT. We hope our findings provide deeper insights into the transferability of LLMs at scale for the research community.</abstract>
    <identifier type="citekey">zheng-etal-2024-breaking</identifier>
    <location>
        <url>https://aclanthology.org/2024.emnlp-main.441</url>
    </location>
    <part>
        <date>2024-11</date>
        <extent unit="page">
            <start>7725</start>
            <end>7738</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale
%A Zheng, Wenzhen
%A Pan, Wenbo
%A Xu, Xu
%A Qin, Libo
%A Yue, Li
%A Zhou, Ming
%Y Al-Onaizan, Yaser
%Y Bansal, Mohit
%Y Chen, Yun-Nung
%S Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
%D 2024
%8 November
%I Association for Computational Linguistics
%C Miami, Florida, USA
%F zheng-etal-2024-breaking
%X In recent years, Large Language Models (LLMs) have made significant strides towards Artificial General Intelligence. However, training these models from scratch requires substantial computational resources and vast amounts of text data. In this paper, we explores an alternative approach to constructing a LLM for a new language by continually pre-training (CPT) from existing pre-trained LLMs, instead of using randomly initialized parameters. Based on parallel experiments on 40 model sizes ranging from 40M to 5B parameters, we find that 1) CPT converges faster and saves significant resources in a scalable manner. 2) CPT adheres to an extended scaling law derived from with a joint data-parameter scaling term. 3) The compute-optimal data-parameter allocation for CPT markedly differs based on our estimated scaling factors. 4) The effectiveness of transfer scale is influenced by training duration and linguistic properties, while robust to data replaying, a method that effectively mitigates catastrophic forgetting in CPT. We hope our findings provide deeper insights into the transferability of LLMs at scale for the research community.
%U https://aclanthology.org/2024.emnlp-main.441
%P 7725-7738

Markdown (Informal)

[Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale](https://aclanthology.org/2024.emnlp-main.441) (Zheng et al., EMNLP 2024)

• Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale (Zheng et al., EMNLP 2024)

ACL

• Wenzhen Zheng, Wenbo Pan, Xu Xu, Libo Qin, Li Yue, and Ming Zhou. 2024. Breaking Language Barriers: Cross-Lingual Continual Pre-Training at Scale. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 7725–7738, Miami, Florida, USA. Association for Computational Linguistics.