邓娟：TG2工作后续计划-10页_1mb

6G Network In-Network AI Architecture Work Plan and Progress Summary (Focusing on 2023-2024)

1. Overview

• TG2 aims to define the foundational architecture for communication-computation fusion networks, build the NetworkAI framework, and assess key enabling technologies and relevant standards organizations.
• The research covers identifying technical characteristics of 6G in-network AI, evaluating their impact on network architecture, and influencing standardization efforts.

2. Historical Work (2021-2023)

• 2021: Identified basic technical features of 6G in-network AI; released the "10 Questions on In-Network AI" white paper.
• 2022: Explored use cases for 6G wireless in-network AI architecture; released "Focus and Analysis of Typical Use Cases" white paper.
• 2023: Proposed multiple 6G in-network AI architectures and key technologies; published series of reports like "RAN Architecture."

3. Current Focus and 2024 Objectives

• Current Priorities: Architecture design enabled by AI.
• 2024 Goals: Complete the design of the 6G wireless in-network AI architecture (v10), building on existing work for international influence.

4. Key Research Areas

• Definition of the 6G in-network AI architecture framework, including logical, data, control, computational, and management layers.
• Challenges:
  • Fragmented consensus among minimal study areas (e.g., QoAIS assessment mechanisms, AI lifecycle management, computational integration/tensor fusion support).
  • Performance optimization space, high uncertainty, and low industry awareness of current solutions.

5. Technical Work Plan (Future Milestones)

• 2024 H1: Complete architecture "convergence," refine v10, conduct technical deep dives into areas like QoAIS, resource allocation, computation integration, and immersive user experiences.
• 2024 H2: Refine to v20, advance solutions for autonomous networks, AI-enabled computation, and multi-agent collaboration.
• Key Milestones:
  • Release QoAIS assessment and guarantee technology; develop infrastructure for autonomous operations/extended reality (XR).
  • Continue technical tracking with groups focusing on frontiers such as large models and their impact on AI-6G fusion; expand analyses for international cooperation.

6. Strategy and Execution Approach

• Organization Method: Task-driven, small-team driven collaborative model to ensure efficiency and expertise utilization for convergence.
• Key Approaches:
  • Continuously refine through internal TG2 discussions and international symposia.
  • Promote results via white papers, prototypes, academic conferences, exhibitions, and participation in national projects.
  • Establish a publication, presentation and dissemination plan (proceedings, white papers, exhibits, telecommunication prototypes).

7. Key Proposals and Actions

• Consolidation: TG2 plans to align with the internal 6GANA TG through multiple rounds of structured discussions to produce standardized and innovative outputs such as the Final Architecture v10 and Milestone-based v20.
• Group Structure: QoAIS Group, Computing and Computation Fusion Group, Network Autonomy Group, Immersive Experience Group, among others.

8. Concluding Remarks

Best efficiency and scope shall be achieved through focused teamwork across companies and via internationally coordinated technological research and industry unification.