The data article presents the large bilingual parallel corpus of
low-resourced language pair Sanskrit-Hindi, named SAHAAYAK 2023. The corpus
contains total of 1.5M sentence pairs between Sanskrit and Hindi. To make the
universal usability of the corpus and to make it balanced, data from multiple
domain has been incorporated into the corpus that includes, News, Daily
conversations, Politics, History, Sport, and Ancient Indian Literature. The
multifaceted approach has been adapted to make a sizable multi-domain corpus of
low-resourced languages like Sanskrit. Our development approach is spanned from
creating a small hand-crafted dataset to applying a wide range of mining,
cleaning, and verification. We have used the three-fold process of mining:
mining from machine-readable sources, mining from non-machine readable sources,
and collation from existing corpora sources. Post mining, the dedicated
pipeline for normalization, alignment, and corpus cleaning is developed and
applied to the corpus to make it ready to use on machine translation
algorithms.

本文介绍了一个名为 SAHAAYAK 2023 的低资源语言 Sanskrit-Hindi 的大型双语平行语料库，包括新闻，日常对话，政治，历史，体育和古印度文学的多个领域，以及通过挖掘，清洗和验证而得到的多方面的方法制作的一个规模可观的多领域语料库，并应用于机器翻译算法。

SAHAAYAK 2023 - 梵文到印地语的多领域双语平行语料库，用于机器翻译

SAHAAYAK 2023 -- the Multi Domain Bilingual Parallel Corpus of Sanskrit  to Hindi for Machine Translation

Recently, numbers of works shows that the performance of neural machine
translation (NMT) can be improved to a certain extent with using visual
information. However, most of these conclusions are drawn from the analysis of
experimental results based on a limited set of bilingual sentence-image pairs,
such as Multi30K. In these kinds of datasets, the content of one bilingual
parallel sentence pair must be well represented by a manually annotated image,
which is different with the actual translation situation. Some previous works
are proposed to addressed the problem by retrieving images from exiting
sentence-image pairs with topic model. However, because of the limited
collection of sentence-image pairs they used, their image retrieval method is
difficult to deal with the out-of-vocabulary words, and can hardly prove that
visual information enhance NMT rather than the co-occurrence of images and
sentences. In this paper, we propose an open-vocabulary image retrieval methods
to collect descriptive images for bilingual parallel corpus using image search
engine. Next, we propose text-aware attentive visual encoder to filter
incorrectly collected noise images. Experiment results on Multi30K and other
two translation datasets show that our proposed method achieves significant
improvements over strong baselines.

本文提出使用图像搜索引擎和文本感知的注意力视觉编码器来收集并过滤具有描述性的图像，以加强神经机器翻译的性能。在多个数据集上进行的实验证明，该方法较强的基线实现了显著的性能提升。