2 Justification

The present Work Item is a follow up of the WI-0054 and WI-0107, which produced an initial series of developer guides covering basic oneM2M features and capabilities. The present WI proposes a new series of guides focusing on advanced topics from Release 3 and Release 4, particularly those related to AI enablement.

The following trends motivate this work:

Growing demand for AI integration in IoT: The convergence of IoT and AI is accelerating rapidly. Developers building AI-powered applications on top of oneM2M platforms require concrete guidance on how to leverage oneM2M's semantic and resource management capabilities for AI model training and inference.

Large Action Models (LAM) for IoT device control: Recent advances in large language models (LLMs) and LAMs have opened new possibilities for natural language-based device control. However, a critical bottleneck is the lack of structured, semantically rich training data for IoT device operations. oneM2M's SDT and SAREF ontology provide exactly the structured device descriptions needed to address this gap. A developer guide on this topic is essential to bridge AI practitioners and oneM2M IoT specialists.

Federated Learning (FL) using native oneM2M resources: FL is increasingly adopted for privacy-preserving distributed model training in IoT deployments. oneM2M's hierarchical architecture (IN-CSE, MN-CSE, ASN-CSE) naturally maps to the FL topology of a central aggregation server and distributed edge training nodes. Using native oneM2M resources (AEs, containers, contentInstances) to manage the full FL lifecycle, without external middleware, is a compelling capability that requires detailed developer guidance.

Continuation of TR-0045 semantic work: TR-0045 provided an introductory guide to SAREF usage in oneM2M. The proposed LAM guide goes significantly further, demonstrating advanced semantic exploitation for AI training data generation, which is not covered in any existing oneM2M TR.