Hugging Face's logo

Datasets:
ruscorpora
/
rnc_enru_aligned

Tasks:
Translation
Other
Languages:
English
Russian
Size:
n<1K
Tags:
word-alignment
manual-annotation
alignment
License:
Dataset card Files
xet
Community
Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

RNC EnRu Aligned (Russian-English Word Alignment Dataset)

Dataset Description

This dataset contains sentence pairs with manual word alignment annotations. It is built upon the Russian-English subcorpus of the Russian National Corpus (RNC).

The primary purpose of this dataset is to support word alignment tasks. The annotation distinguishes between "Sure" and "Possible" alignment links, following standard alignment conventions.

  • Curated by: Russian National Corpus
  • Language(s): English, Russian
  • License: CC BY-NC-SA 4.0

Dataset Statistics

  • Current Size: 881 sentence pairs.
  • Status: Active (The dataset is currently being expanded).

Dataset Structure

The dataset is organized into three directories representing the full set and the training/validation splits.

Directory Layout

  • all/: Contains the complete dataset (all available annotated pairs).
  • train/: Contains 90% of the data from all, intended for training.
  • dev/: Contains the remaining 10% of the data, intended for development/validation.

File Format

Each directory contains three parallel files sharing the prefix enru:

  1. enru.src: Source sentences in English. One sentence per line.
  2. enru.tgt: Target sentences in Russian. One sentence per line.
  3. enru.talp: Word alignment annotations in Pharaoh format.

The .talp files contain the alignment links for the corresponding sentence pairs.

  • Notation: Space-separated pairs of indices i-j (e.g., 1-1 2p3).
  • Indexing: 1-based indexing (The first word in the sentence has index 1, not 0).
  • Link Types: The dataset contains both S (Sure) and P (Possible) links.

Dataset Creation

Preprocessing

Both languages were tokenized with punctuation treated as separate tokens.

  • Russian: Hyphenated words (e.g., кое-кто) remain single tokens.
  • English: Contractions like n't and possessives like 's are split into separate tokens. Won't is normalized to will + not.

Annotation Process

Three native Russian speakers (English level B2+) performed manual annotation using Sure (S) and Possible (P) labels, based on principles from [Politova et al., 2023]. Language-specific Annotation Rules:

  • Articles: Not aligned.
  • Auxiliaries & Prepositions: Aligned as S if an explicit counterpart exists; otherwise marked as P (attached to the main word) or left unaligned.
  • Negation: n't aligns with не (S).
  • Idioms: Marked as S for stable translations, otherwise P.

Quality Control

Annotations underwent selective joint verification by all annotators to ensure consistency, though parallel double-annotation was not performed.

License

This dataset is distributed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.

How to cite

If you are using this dataset in your scientific work, please, cite this paper:

TBA

Downloads last month
8