Go to ToC          Scope->

1        Introduction

(Tentative)

MPAI, the international, unaffiliated, and not-for-profit organisation developing standard for AI-based data coding [1] established in 2020, has developed a range of Technical Specifications and Technical Reports [2] – 18 and 2, respectively – covering a variety of application domains that can benefit from the use of Artificial Intelligence (AI).

Two standards relevant to this document are

  1. Technical Specification: AI Framework (MPAI-AIF) [11] specifying initialisation, dynamic configuration, and control of AI applications implemented as AI Workflows (AIW) composed of AI Modules (AIM). An overview and the full text of the standard are available.
  2. Technical Specification: MPAI Metaverse Model (MPAI-MMM) – Technologies (MMM-TEC) [12]. An overview and the full text of the standard are available.

MMM-TEC assumes that

  1. M-Instances are metaverse instances implemented in conformity with MMM-TEC.
  2. An M-Instance is populated with Processes acting on Items representing “things” in the M-Instance.
  3. User is a Process operating under the responsibility of a human, either
    1. Directly – Human User (H-User), or
    2. Without direct control – Autonomous User (A-User).

MPAI believes that

  1. A-Users able to interact with Users – both A- and H-Users – are attractive components for M-Instances.
  2. A component-based A-User standard can accelerate the development and deployment at scale of A-Users.
  3. MPAI has the necessary capabilities and technologies to start a new activity called Pursuing Goals in metaverse (MPAI-PGM).

Therefore, MPAI has initiated the MPAI-PGM standard project called Autonomous User Architecture (PGM-AUA). Use Cases and Functional Requirements: Pursuing Goals in metaverse (MPAI-PGM) – Autonomous User Architecture (PGM-AUA) [6] develops A-Users functional requirements based on an initial set of Use Cases. Outstanding among them is that A-Users should be implementable with technologies that do not require high levels of processing and energy consumption. For example, some language and reasoning abilities are obviously required but an A-User does not necessarily require the use of general-purpose Large Language Models.

A-Users require a large set of technologies that can be made available by AI Modules integrated in an AI Workflow operating in the Standard AI Framework [11]. MPAI intends to re-use its considerable set of technologies developed in the contexts of its standards – most of them based on AI-Framework – when relevant to PGM-AUA and to obtain those not yet available by publishing the Call for Technologies: Pursuing Goals in metaverse (MPAI-PGM) – Autonomous User Architecture (PGM-AUA) [6]]. The present document – an attachment to PGM-AUA Use Cases and Functional Requirements – is a non-official PGM-AUA Technical Specification and is at most described by the adjective tentative that Respondents to the Call may either consider and comment on in their Responses or not consider at all.

Go to ToC          Scope->