R1-2410050
discussion
Discussion on specification support for CSI prediction
From Fujitsu
Summary
Fujitsu presents evaluation results and proposals for ensuring consistency between training and inference for AI/ML-based CSI prediction with UE-side models in 5G NR. The document contains 1 observation and 2 proposals focused on identifying key network-side conditions that impact model generalization performance.
Position
Fujitsu advocates FOR using associated ID as the primary mechanism to ensure training/inference consistency for CSI prediction, similar to the approach already agreed for beam management. They are pushing AGAINST considering tilt angle as a key network-side condition, based on their simulation results showing negligible performance impact. They support focusing on gNB antenna configurations and scenario types (indoor/outdoor, LOS/NLOS) as the critical factors that should be associated with the same ID.
Key proposals
- Observation 1 (Sec 2.1): For AI/ML based CSI prediction, the tilt angle has negligible impact on consistency between training and inference
- Proposal 1 (Sec 2.1): The tilt angle is not the key aspect of NW-side additional conditions to ensure consistency between training and inference
- Proposal 2 (Sec 2.2): Regarding NW-side conditions, RAN1 to consider associated ID as starting point to address consistency between training and inference for CSI prediction with UE side model. With the same associated ID, the UE could assume the same gNB antenna configurations, and/or the same scenario (e.g., indoor/outdoor, LOS/NLOS) of the cell. RAN1 to further discuss how to apply the associated ID across different cells.