(Informative)
Technical Specification: AI Framework (MPAI-AIF) V2.1 is a key element of the MPAI approach to AI-based Data Coding standards. It is based on a framework enabling initialisation, dynamic configuration, and control of AIWs in the standard AI Framework environment depicted in Figure 1. The Data Data produced by executing specific functions by AI Modules (AIM) are communicated to other AIMs in an AIW.
The functions performed by an AIM may improve if it knows more about the capabilities of the AIMs it is connected to and the Data they receive. For example, an instance of the MPAI Natural Language Processing (MMC-NLU) AIM has the task to refine the text it receives and produce the Meaning of the Text. This can be dome using other sources of information, such as:
- The identifiers of the object referenced in the text.
- The context of the object in a relevant space.
If the instance of the NLU AIM has access to this additional information, it is likely that an AIM able to process it will provide improved accuracy of the refined text and Meaning.
Technical Specification: AI Module Profiles (MPAI-PRF) enables an AIM instance to signal the Attributes – that uniquely characterise it, e.g. input data, output data, and functionality and Sub-Attributes – such as languages supported by a Text and Speech Translation AIM. Currently, MPAI-PRF defines the Attributes of eight AIMs but Profiles for more AIMs are likely to be defined in the future.
The effectiveness of the functions performed by an AIM can be enabled or enhanced if the AIM has even more knowledge about the characteristics of the Data received. Examples of characteristics include:
- The CIE 1931 colour space of an instance of the Visual Data Type.
- The MP3 format of a speech segment.
- The WAV file format of an audio segment.
- The gamma correction applied to the device that produced a video.
- The Instance ID of an object in an audio segment.
- The Text conveyed by a speech segment.
Technical Specification: Data Types, Formats, and Attributes (MPAI-TFA) V1.3 specifies the Qualifier Data Type, a container that can be used to represent, for instance, that a Visual Data Type instance:
- Uses a given colour space (Sub-Type)
- Was produced by an AVC encoder (Format).
- Is described by Dublin Core Metadata (Attribute).
Therefore, Qualifiers are a specialised type of metadata intended to support the operation of AIMs receiving data from other AIMs and conveying information on Sub-Types, Formats, and Attributes related to the Content. The information conveyed by Qualifiers is intended for use by an AIM, even though they are human-readable. The combination of “Content” (the Data of a Data Type) and “Qualifier” (the combination of Sub-Type, Format, and Attributes) is called “Object“.
MPAI provides a standard method to attach information to a Data Type instance called Annotation. This is defined as Data attached to an Object or a Scene. As opposed to a Qualifier that describes the intrinsic properties of a Data Type, an Annotation is spatially and temporally local and changeable.
MPAI plans of publishing new versions of MPAI-TFA each time an application standard requires Qualifiers or when there is a need to extend existing Qualifiers. MPAI-TFA users may communicate their need for extension of existing and specification of additional Data Type Qualifiers to the MPAI Secretariat. Therefore, versioning of Qualifiers is a critical component of MPAI-TFA.
The Chapters, Sections, and Annexes of this Technical Specification are Normative unless they are explicitly labelled as Informative. In all Chapters and Sections, Terms beginning with a capital letter are defined in Table 1 if they are specific to this Technical Specification. All MPAI-defined Terms are accessible online. All Chapters and Annexes are Normative unless they are labelled as Informative.
