Research

Paper

TESTING March 25, 2026

Samasāmayik: A Parallel Dataset for Hindi-Sanskrit Machine Translation

Authors

N J Karthika, Keerthana Suryanarayanan, Jahanvi Purohit, Ganesh Ramakrishnan, Jitin Singla, Anil Kumar Gourishetty

Abstract

We release Samasāmayik, a novel, meticulously curated, large-scale Hindi-Sanskrit corpus, comprising 92,196 parallel sentences. Unlike most data available in Sanskrit, which focuses on classical era text and poetry, this corpus aggregates data from diverse sources covering contemporary materials, including spoken tutorials, children's magazines, radio conversations, and instruction materials. We benchmark this new dataset by fine-tuning three complementary models - ByT5, NLLB and IndicTrans-v2, to demonstrate its utility. Our experiments demonstrate that models trained on the Samasamayik corpus achieve significant performance gains on in-domain test data, while achieving comparable performance on other widely used test sets, establishing a strong new performance baseline for contemporary Hindi-Sanskrit translation. Furthermore, a comparative analysis against existing corpora reveals minimal semantic and lexical overlap, confirming the novelty and non-redundancy of our dataset as a robust new resource for low-resource Indic language MT.

Metadata

arXiv ID: 2603.24307
Provider: ARXIV
Primary Category: cs.CL
Published: 2026-03-25
Fetched: 2026-03-26 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.24307v1</id>\n    <title>Samasāmayik: A Parallel Dataset for Hindi-Sanskrit Machine Translation</title>\n    <updated>2026-03-25T13:48:34Z</updated>\n    <link href='https://arxiv.org/abs/2603.24307v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.24307v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We release Samasāmayik, a novel, meticulously curated, large-scale Hindi-Sanskrit corpus, comprising 92,196 parallel sentences. Unlike most data available in Sanskrit, which focuses on classical era text and poetry, this corpus aggregates data from diverse sources covering contemporary materials, including spoken tutorials, children's magazines, radio conversations, and instruction materials. We benchmark this new dataset by fine-tuning three complementary models - ByT5, NLLB and IndicTrans-v2, to demonstrate its utility. Our experiments demonstrate that models trained on the Samasamayik corpus achieve significant performance gains on in-domain test data, while achieving comparable performance on other widely used test sets, establishing a strong new performance baseline for contemporary Hindi-Sanskrit translation. Furthermore, a comparative analysis against existing corpora reveals minimal semantic and lexical overlap, confirming the novelty and non-redundancy of our dataset as a robust new resource for low-resource Indic language MT.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-25T13:48:34Z</published>\n    <arxiv:primary_category term='cs.CL'/>\n    <author>\n      <name>N J Karthika</name>\n    </author>\n    <author>\n      <name>Keerthana Suryanarayanan</name>\n    </author>\n    <author>\n      <name>Jahanvi Purohit</name>\n    </author>\n    <author>\n      <name>Ganesh Ramakrishnan</name>\n    </author>\n    <author>\n      <name>Jitin Singla</name>\n    </author>\n    <author>\n      <name>Anil Kumar Gourishetty</name>\n    </author>\n  </entry>"
}