Go To MPAI-MMC AI Modules

1     Function 2     Reference Model 3     Input/Output Data
4     SubAIMs 5     JSON Metadata 6     Profiles
7     Reference Software 8     Conformance Texting 9     Performance Assessment

1     Functions

Text-to-Text Translation:

Receives Selector Determining the input and target language.
Text Object Text to be translated.
Meaning Input Text Meaning.
Produces Translated Text Output Translates Text.

2     Reference Model

Figure 1 depicts the Reference Model of the Text-to-Text Translation AIM.

Figure 1 – Text-to-Text Translation AIM Reference Model

3    Input/Output Data

Table 1 specifies the Input and Output Data of the Text-to-Text Translation AIM.

Table 1 – I/O Data of the Text-to-Text Translation AIM

Input Description
Text Object Input Text Object.
Meaning Meaning of Input Text
Language Selector Input and target Language.
Output Description
Translated Text Translation of Text (or Refined  Text).

4     SubAIMs

No SubAIMs.

5     JSON Metadata

https://schemas.mpai.community/MMC/V2.3/AIMs/TextToTextTranslation.json

6     Profiles

No Profiles.

7     Reference Software

8     Conformance Testing

Table 2 provides the Conformance Testing Method for MMC-TTT AIM.

If a schema contains references to other schemas, conformance of data for the primary schema implies that any data referencing a secondary schema shall also validate against the relevant schema, if present and conform with the Qualifier, if present.

Table 2 – Conformance Testing Method for MMC-TTT AIM

Input Language Selector Shall validate against Language Selector schema.
Text Object Shall validate against Text Object schema.
Text Data shall conform with Text Qualifier.
Meaning Shall validate against Meaning schema.
Output Translated Text Object Shall validate against Text Object schema.
Text Data shall conform with Text Qualifier.

Table 3 provides an example of MMC-TTT AIM conformance testing.

Table 3 – An example MMC-TTT AIM conformance testing

Input Data Data Type Input Conformance Testing Data
Input Text Unicode All input Text files to be drawn from Text files.
Output Data Data Type Output Conformance Testing Criteria
Translated Text Unicode All Text files produced shall conform with Text.

9     Performance Assessment

Performance Assessment of an MMC-TTT AIM Implementation shall be performed using a dataset of text sentences in a given language. Each text sentence shall have at least one translated text.

The Performance Assessment Report of an MMC-TTT AIM Implementation shall include:

  1. The Identifier of the MMC-TTT AIM.
  2. The Identifier of the dataset of text sentences.
  3. The name of the input and output languages and their ISO 639 Set 3 three-letter code.
  4. The number of text sentences in the  data set and the average number of translated texts per input text.
  5. The maximum value N of n-grams used.
  6. The BLEU Score of the MMC-TTT AIM, defined as the Arithmetic Mean of the individual BLEU Scores computed over the dataset, where each BLEU Score is the product of the Brevity Penalty and the Geometric Mean Precision, and where:
    1. The Brevity Penalty of a candidate translation of length c to a reference translation of length r is min (1,e(1-r/c)).
    2. The Sentence Precision of a set of N n-grams is exp(∑n=1,Nlog(pn)/N, where pi is the precision of the i-th n-gram.

Go To MPAI-MMC AI Modules