R1-2409628 discussion

Discussion on other aspects of AI/ML model and data

From Spreadtrum
Status: not treated
WI: NR_AIML_air
Agenda: 9.1.4.2
Release: Rel-19
Source: 3gpp.org ↗
Spreadtrum's prior position on 9.1.4.2 at RAN1#118bis · AI-synthesized, paraphrased
verify sources →
Advocates for deprioritizing model transfer cases z1 and z2 in Rel-19 due to cross-vendor collaboration burdens, supports mechanism 1a for UE data collection to avoid privacy exposure, and promotes network-controlled model identification.

Summary

Spreadtrum presents three proposals and four observations regarding AI/ML for the NR air interface in Rel-19, focusing on data collection, model transfer, and identification. The document argues that RAN1 should deprioritize certain model transfer cases and exclude specific model identification options for two-sided models, while deferring data collection discussions to RAN2. It emphasizes that RAN1's scope should remain on model structure content rather than signaling formats.

Position

Spreadtrum proposes that RAN1 should not discuss data collection for UE-side model training, preferring mechanism 1a or waiting for RAN2 progress due to privacy concerns regarding network exposure. They require that model transfer/delivery Case z1 be deprioritized in Rel-19, arguing it incurs unnecessary offline cross-vendor collaboration burdens compared to Case y. They oppose considering MI-Option 1 for two-sided use cases, asserting that the data collection steps (A-C) are unnecessary for model identification and that Step E can be realized by MI-Option 2 or 3. They observe that RAN1 should focus on the content of model structure, leaving model format and signaling to other working groups. They support considering MI-Option 2, 3, and 4 for two-sided use cases, noting that MI-Option 4 depends on defining reference models.

Key proposals

Your notes

Private to your account