@inproceedings{nehrdich-etal-2026-mitrasamgraha,
title = "Mitrasamgraha: A Comprehensive Classical {S}anskrit Machine Translation Dataset",
author = "Nehrdich, Sebastian and
Allport, David and
Sandhan, Jivnesh and
Jagadeeshan, Manoj Balaji and
Kumar, Sujeet and
Sellmer, Sven and
Goyal, Pawan and
Keutzer, Kurt",
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Eleventh Conference on Machine Translation",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.wmt-main.27/",
pages = "507--519",
ISBN = "979-8-89176-474-3",
abstract = "Although machine translation is often considered solved for high-resource languages, it remains challenging for texts involving poetic language, philosophical concepts, and layered metaphors, characteristics that are central to Sanskrit literature. Additionally, Sanskrit{'}s rich morphology, sandhi, and compounding, combined with its multi-millennial and multi-domain textual tradition, highlight the need for large, diverse parallel resources, which are currently scarce. We introduce Mitrasaṃgraha, a large-scale Sanskrit-English machine translation dataset containing 391,548 aligned sentence pairs, over 4 times larger than the previously available Itih{\={a}}sa corpus (Aralikatte et al., 2021). The dataset spans more than 3 millennia of Sanskrit literature across multiple domains and includes temporally annotated data for studying domain and period effects on translation performance. We release manually corrected development (5,587) and test (5,552) sets. Benchmarks show that fine-tuning models such as NLLB and Gemma substantially improves translation quality, and retrieval-augmented prompting further enhances performance. In a controlled cross-dataset comparison, training on Mitrasaṃgraha outperforms both Itih{\={a}}sa and web-mined bitext for two model families. Despite these gains, our results reveal persistent challenges in translating complex compounds, philosophical terminology, and metaphorical language in Sanskrit, highlighting the need for further research in this area."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nehrdich-etal-2026-mitrasamgraha">
<titleInfo>
<title>Mitrasamgraha: A Comprehensive Classical Sanskrit Machine Translation Dataset</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sebastian</namePart>
<namePart type="family">Nehrdich</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Allport</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jivnesh</namePart>
<namePart type="family">Sandhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Manoj</namePart>
<namePart type="given">Balaji</namePart>
<namePart type="family">Jagadeeshan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sujeet</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sven</namePart>
<namePart type="family">Sellmer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pawan</namePart>
<namePart type="family">Goyal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kurt</namePart>
<namePart type="family">Keutzer</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 Eleventh Conference on Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Barry</namePart>
<namePart type="family">Haddow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Kocmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philipp</namePart>
<namePart type="family">Koehn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christof</namePart>
<namePart type="family">Monz</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">979-8-89176-474-3</identifier>
</relatedItem>
<abstract>Although machine translation is often considered solved for high-resource languages, it remains challenging for texts involving poetic language, philosophical concepts, and layered metaphors, characteristics that are central to Sanskrit literature. Additionally, Sanskrit’s rich morphology, sandhi, and compounding, combined with its multi-millennial and multi-domain textual tradition, highlight the need for large, diverse parallel resources, which are currently scarce. We introduce Mitrasaṃgraha, a large-scale Sanskrit-English machine translation dataset containing 391,548 aligned sentence pairs, over 4 times larger than the previously available Itihāsa corpus (Aralikatte et al., 2021). The dataset spans more than 3 millennia of Sanskrit literature across multiple domains and includes temporally annotated data for studying domain and period effects on translation performance. We release manually corrected development (5,587) and test (5,552) sets. Benchmarks show that fine-tuning models such as NLLB and Gemma substantially improves translation quality, and retrieval-augmented prompting further enhances performance. In a controlled cross-dataset comparison, training on Mitrasaṃgraha outperforms both Itihāsa and web-mined bitext for two model families. Despite these gains, our results reveal persistent challenges in translating complex compounds, philosophical terminology, and metaphorical language in Sanskrit, highlighting the need for further research in this area.</abstract>
<identifier type="citekey">nehrdich-etal-2026-mitrasamgraha</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.27/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>507</start>
<end>519</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Mitrasamgraha: A Comprehensive Classical Sanskrit Machine Translation Dataset
%A Nehrdich, Sebastian
%A Allport, David
%A Sandhan, Jivnesh
%A Jagadeeshan, Manoj Balaji
%A Kumar, Sujeet
%A Sellmer, Sven
%A Goyal, Pawan
%A Keutzer, Kurt
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Eleventh Conference on Machine Translation
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-474-3
%F nehrdich-etal-2026-mitrasamgraha
%X Although machine translation is often considered solved for high-resource languages, it remains challenging for texts involving poetic language, philosophical concepts, and layered metaphors, characteristics that are central to Sanskrit literature. Additionally, Sanskrit’s rich morphology, sandhi, and compounding, combined with its multi-millennial and multi-domain textual tradition, highlight the need for large, diverse parallel resources, which are currently scarce. We introduce Mitrasaṃgraha, a large-scale Sanskrit-English machine translation dataset containing 391,548 aligned sentence pairs, over 4 times larger than the previously available Itihāsa corpus (Aralikatte et al., 2021). The dataset spans more than 3 millennia of Sanskrit literature across multiple domains and includes temporally annotated data for studying domain and period effects on translation performance. We release manually corrected development (5,587) and test (5,552) sets. Benchmarks show that fine-tuning models such as NLLB and Gemma substantially improves translation quality, and retrieval-augmented prompting further enhances performance. In a controlled cross-dataset comparison, training on Mitrasaṃgraha outperforms both Itihāsa and web-mined bitext for two model families. Despite these gains, our results reveal persistent challenges in translating complex compounds, philosophical terminology, and metaphorical language in Sanskrit, highlighting the need for further research in this area.
%U https://aclanthology.org/2026.wmt-main.27/
%P 507-519
Markdown (Informal)
[Mitrasamgraha: A Comprehensive Classical Sanskrit Machine Translation Dataset](https://aclanthology.org/2026.wmt-main.27/) (Nehrdich et al., WMT 2026)
ACL
- Sebastian Nehrdich, David Allport, Jivnesh Sandhan, Manoj Balaji Jagadeeshan, Sujeet Kumar, Sven Sellmer, Pawan Goyal, and Kurt Keutzer. 2026. Mitrasamgraha: A Comprehensive Classical Sanskrit Machine Translation Dataset. In Proceedings of the Eleventh Conference on Machine Translation, pages 507–519, Budapest, Hungary. Association for Computational Linguistics.