mqm.yml

template_name: mqm
template_label: MQM
template_description: A variation proposed by Freitag et al., 2021 of the Multidimensional Quality Metrics

# ========================================================================================
#
# The original guidelines for MQM are detailed here: https://aclanthology.org/2013.tc-1.6
# The full data collection for this typology, as well as the original web viewer for MQM are here:
# https://github.com/google/wmt-mqm-human-evaluation
#
# ========================================================================================

interface_text:
  typology:
    source_label: "Original Sentence"
    target_label: "German Translation"

edits:
  - name: accuracy
    label: "Accuracy"
    color: red
    icon: fa-check
    enable_output: true  
    annotation:
      - name: accuracy_type
        question: "Select the accuracy problem."
        options:
          - name: addition
            label: "Addition"
            question: "Rate the severity"
            options: likert-3
          - name: mistranslation
            label: "Mistranslation"
            questions: "Rate the severity"
            options: likert-3
          - name: omission
            label: "Omission"
            questions: "Rate the severity"
            options: likert-3
          - name: untranslated_text
            label: "Untranslated text"
            questions: "Rate the severity"
            options: likert-3
  - name: fluency
    label: "Fluency"
    color: teal
    icon: fa-comment-slash
    enable_output: true   
    annotation:
      - name: fluency_type
        question: "Select the fluency problem."
        options:
          - name: grammar
            label: "Grammar"
            question: "Rate the severity"
            options: likert-3
          - name: inconsistency
            label: "Inconsistency"
            questions: "Rate the severity"
            options: likert-3
          - name: punctuation
            label: "Punctuation"
            questions: "Rate the severity"
            options: likert-3
          - name: register
            label: "Register"
            questions: "Rate the severity"
            options: likert-3
          - name: spelling
            label: "Spelling"
            questions: "Rate the severity"
            options: likert-3  
  - name: locale_convention
    label: "Locale Convention"
    color: yellow
    icon: fa-people-arrows 
    enable_output: true   
    annotation:
      - name: locale_convention_type
        question: "Select the convention problem."
        options:
          - name: address_format
            label: "Address format"
            question: "Rate the severity"
            options: likert-3
          - name: currency_format
            label: "Currency format"
            questions: "Rate the severity"
            options: likert-3
          - name: date_format
            label: "Date format"
            questions: "Rate the severity"
            options: likert-3
          - name: time_format
            label: "Time format"
            questions: "Rate the severity"
            options: likert-3 
  - name: akward
    label: "Style"
    color: orange
    icon: fa-person-drowning 
    enable_output: true   
    annotation:
      - name: akward_type
        label: "Severity"
        question: "Rate the severity."
        options: likert-3
  - name: terminology
    label: "Terminology"
    color: green
    icon: fa-flask
    enable_output: true   
    annotation:
      - name: terminology_type
        question: "Select the Terminology problem."
        options:
          - name: inappropriate_for_context
            label: "Inappropriate for context"
            question: "Rate the severity"
            options: likert-3
          - name: inconsistent_use_of_terminology
            label: "Inconsistent use of terminology"
            questions: "Rate the severity"
            options: likert-3
  - name: non_translation
    label: "Non-translation"
    color: blue
    icon: fa-xmark
    enable_output: true   
    annotation:
      - name: non_translation_type
        label: "Severity"
        question: "Rate the severity."
        options: likert-3
  - name: other
    label: "Other"
    color: grey
    icon: fa-ellipsis
    enable_output: true   
    annotation:
      - name: other_type
        label: "Severity"
        question: "Rate the severity."
        options: likert-3
paper_link: https://aclanthology.org/2021.tacl-1.87
citation: |
  @article{freitag-etal-2021-experts,
    title = "Experts, Errors, and Context: A Large-Scale Study of Human Evaluation for Machine Translation",
    author = "Freitag, Markus  and
      Foster, George  and
      Grangier, David  and
      Ratnakar, Viresh  and
      Tan, Qijun  and
      Macherey, Wolfgang",
    journal = "Transactions of the Association for Computational Linguistics",
    volume = "9",
    year = "2021",
    address = "Cambridge, MA",
    publisher = "MIT Press",
    url = "https://aclanthology.org/2021.tacl-1.87",
    doi = "10.1162/tacl_a_00437",
    pages = "1460--1474",
  }