S4-260185 - AI Proposals

[AI_IMS_MED] Call flow for split inferencing loop

Back to Agenda Download Proposals
AI-Generated Proposals AI

Proposal: Add the following text to the base CR:

The UE and the DCAS (via the MF) configure the exchanged intermediate data format parameters over ADC

The intermediate data format parameters may include information on tensors characteristics, including the compression profile identifiers.

The UE captures input media data and provides it to inference task running on the UE.

The UE executes the UE inference subtask on the input media data using selected UE submodel and generates intermediate data for continuation of the inference process at the DCAS.

The UE transmits the intermediate data to DCAS (Via MF) according to the selected intermediate data format.

DCAS executes inference task on the received intermediate data using selected Remote submodel and generates the processed media data based on the inference results.

DCAS transmits the processed media data to UE (via MF).

The UE renders the final processed media data.

Document Information
Source:
InterDigital Finland Oy
Type:
discussion
For:
Agreement
Original Document:
View on 3GPP
Title: [AI_IMS_MED] Call flow for split inferencing loop
Agenda item: 10.5
Agenda item description: AI_IMS-MED (Media aspects for AI/ML in IMS services)
Doc type: discussion
For action: Agreement
Release: Rel-20
Contact: Stephane Onno
Uploaded: 2026-02-03T19:11:23.097000
Contact ID: 84864
TDoc Status: merged
Reservation date: 03/02/2026 16:35:57
Agenda item sort order: 52