This paper describes additional aspects of a digital tool called the 'Textual
History Tool'. We describe its various salient features with special reference
to those of its features that may help the philologist digitize commentaries
and sub-commentaries on a text. This tool captures the historical evolution of
a text through various temporal stages, and interrelated data culled from
various types of related texts. We use the text of the Kāśikāvrtti (KV)
as a sample text, and with the help of philologists, we digitize the
commentaries available to us. We digitize the Nyāsa (Ny), the Padamañjarī
(Pm) and sub commentaries on the KV text known as the Tantrapradīpa (Tp), and
the Makaranda (Mk). We divide each commentary and sub-commentary into
functional units and describe the methodology and motivation behind the
functional unit division. Our functional unit division helps generate more
accurate phylogenetic trees for the text, based on distance methods using the
data entered in the tool.

本文介绍一种名为 'Textual History Tool' 的数字工具，其主要应用于文献学家数字化文本的注释和子注释，并将其分成功能单元以生成更准确的进化树。

有效数字化注释和子注释的策略：走向文本史构建

Strategies of Effective Digitization of Commentaries and Sub-commentaries: Towards the Construction of Textual History

Effective training of deep neural networks can be challenging, and there
remain many open questions on how to best learn these models. Recently
developed methods to improve neural network training examine teaching:
providing learned information during the training process to improve downstream
model performance. In this paper, we take steps towards extending the scope of
teaching. We propose a flexible teaching framework using commentaries, learned
meta-information helpful for training on a particular task. We present
gradient-based methods to learn commentaries, leveraging recent work on
implicit differentiation for scalability. We explore diverse applications of
commentaries, from weighting training examples, to parameterising
label-dependent data augmentation policies, to representing attention masks
that highlight salient image regions. We find that commentaries can improve
training speed and/or performance, and provide insights about the dataset and
training process. We also observe that commentaries generalise: they can be
reused when training new models to obtain performance benefits, suggesting a
use-case where commentaries are stored with a dataset and leveraged in future
for improved model training.

提出了一种基于评论的灵活教学框架，使用基于梯度的方法来学习有助于模型训练的元信息，并从权重训练样本、参数化标签相关数据增强策略到表示突出显示显著图像区域的注意力掩码等多个方面探讨了评论的广泛应用。结果表明，评论可以提高训练速度和 / 或性能，并提供关于数据集和训练过程的见解。