Skip to content
Showcase/Symlink: Linking Mathematical Symbols to their Descriptions
advancedtext

Symlink: Linking Mathematical Symbols to their Descriptions

SemEval-2022 Task 12 (Symlink): extracting mathematical symbols and their textual descriptions from scientific documents and linking each symbol to the description it refers to. Annotators mark symbol and description spans, then link them with relation types (Direct, Count, Coref-Symbol, Coref-Description). Documents are drawn from arXiv papers across mathematics, physics, biology, economics, and computer science.

Q1: Rate your experience12345Q2: Primary use case?ResearchIndustryEducationQ3: Additional feedback

Configuration Fileconfig.yaml

This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.

yaml
# Symlink: Linking Mathematical Symbols to their Descriptions
# Based on Lai, Pouran Ben Veyseh, Dernoncourt, and Nguyen, SemEval-2022 Task 12
# Paper: https://aclanthology.org/2022.semeval-1.230/
# Dataset/task page: https://sites.google.com/view/symlink
#
# Symlink combines entity recognition and relation extraction over the text
# of scientific documents (from arXiv). Annotators first mark the spans that
# are mathematical symbols and the spans that describe them, then link each
# symbol to its description with one of the task's relation types.
#
# Entity spans:
# - Symbol: a mathematical symbol (e.g., x, \theta, A_{ij})
# - Description: the text that describes/defines a symbol. In the official
#   corpus, descriptions are further distinguished as primary vs. ordered.
#
# Relation types (symbol <-> description links):
# - Direct: a symbol is directly linked to its description
# - Count: a description quantifies a set of symbols
# - Coref-Symbol: two symbol mentions refer to the same entity
# - Coref-Description: two description mentions refer to the same entity

annotation_task_name: "Symlink - Symbols and Descriptions"
task_dir: "."

data_files:
  - sample-data.json

item_properties:
  id_key: "id"
  text_key: "text"

output_annotation_dir: "annotation_output/"
output_annotation_format: "json"

port: 8000
server_name: localhost

annotation_schemes:
  - annotation_type: span
    name: entities
    description: "Mark each mathematical symbol and each description span"
    labels:
      - "Symbol"
      - "Description"

  - annotation_type: radio
    name: relation_type
    description: "For the primary symbol-description pair, what is the relation?"
    labels:
      - "Direct"
      - "Count"
      - "Coref-Symbol"
      - "Coref-Description"
    tooltips:
      "Direct": "A symbol is directly linked to its description"
      "Count": "A description quantifies a set of symbols"
      "Coref-Symbol": "Two symbol mentions refer to the same entity"
      "Coref-Description": "Two description mentions refer to the same entity"

  - annotation_type: text
    name: notes
    description: "Notes on additional symbol-description links in this passage (optional)"
    textarea: true
    required: false

annotation_instructions: |
  Read the passage from a scientific document. First highlight every
  mathematical symbol span and every description span. Then, for the main
  symbol-description pair, select the relation type that links them. Use the
  notes field to record any additional links present in the passage.

html_layout: |
  <div style="padding: 15px; max-width: 820px; margin: auto;">
    <div style="background: #eff6ff; border: 1px solid #bfdbfe; border-radius: 8px; padding: 16px; margin-bottom: 12px;">
      <strong style="color: #1e40af;">Passage:</strong>
      <p style="font-size: 16px; line-height: 1.7; margin: 8px 0 0 0;">{{text}}</p>
    </div>
  </div>

allow_all_users: true
instances_per_annotator: 40
annotation_per_instance: 2
allow_skip: true

Sample Datasample-data.json

json
[
  {
    "id": "sym_001",
    "text": "Let n denote the number of samples in the training set, where n is a positive integer."
  },
  {
    "id": "sym_002",
    "text": "The learning rate alpha controls the step size taken during each gradient update."
  }
]

// ... and 8 more items

Get This Design

View on GitHub

Clone or download from the repository

Quick start:

git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/semeval/2022/task12-symlink
potato start config.yaml

Dataset & paper

Lai et al., SemEval 2022

Citation (BibTeX)

bibtex
@inproceedings{lai-etal-2022-semeval,
    title = "{S}em{E}val 2022 Task 12: Symlink - Linking Mathematical Symbols to their Descriptions",
    author = "Lai, Viet  and Pouran Ben Veyseh, Amir  and Dernoncourt, Franck  and Nguyen, Thien",
    booktitle = "Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)",
    month = jul,
    year = "2022",
    address = "Seattle, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2022.semeval-1.230",
    doi = "10.18653/v1/2022.semeval-1.230",
    pages = "1671--1678"
}

Details

Annotation Types

spanradiotext

Domain

NLPScientific TextSemEval

Use Cases

Entity RecognitionRelation ExtractionScientific Document Understanding

Tags

semevalsemeval-2022shared-taskentity-linkingrelation-extractionmathematicsscientific-text

Found an issue or want to improve this design?

Open an Issue