Rivaling Transformers: Multi-Scale Structured State-Space Mixtures for Agentic 6G O-RAN

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Rezazadeh, Farhad, Chergui, Hatim, Debbah, Merouane, Song, Houbing, Niyato, Dusit, Liu, Lingjia
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914078841110528
author Rezazadeh, Farhad
Chergui, Hatim
Debbah, Merouane
Song, Houbing
Niyato, Dusit
Liu, Lingjia
author_facet Rezazadeh, Farhad
Chergui, Hatim
Debbah, Merouane
Song, Houbing
Niyato, Dusit
Liu, Lingjia
contents In sixth-generation (6G) Open Radio Access Networks (O-RAN), proactive control is preferable. A key open challenge is delivering control-grade predictions within Near-Real-Time (Near-RT) latency and computational constraints under multi-timescale dynamics. We therefore cast RAN Intelligent Controller (RIC) analytics as an agentic perceive-predict xApp that turns noisy, multivariate RAN telemetry into short-horizon per-User Equipment (UE) key performance indicator (KPI) forecasts to drive anticipatory control. In this regard, Transformers are powerful for sequence learning and time-series forecasting, but they are memory-intensive, which limits Near-RT RIC use. Therefore, we need models that maintain accuracy while reducing latency and data movement. To this end, we propose a lightweight Multi-Scale Structured State-Space Mixtures (MS3M) forecaster that mixes HiPPO-LegS kernels to capture multi-timescale radio dynamics. We develop stable discrete state-space models (SSMs) via bilinear (Tustin) discretization and apply their causal impulse responses as per-feature depthwise convolutions. Squeeze-and-Excitation gating dynamically reweights KPI channels as conditions change, and a compact gated channel-mixing layer models cross-feature nonlinearities without Transformer-level cost. The model is KPI-agnostic -- Reference Signal Received Power (RSRP) serves as a canonical use case -- and is trained on sliding windows to predict the immediate next step. Empirical evaluations conducted using our bespoke O-RAN testbed KPI time-series dataset (59,441 windows across 13 KPIs). Crucially for O-RAN constraints, MS3M achieves a 0.057 s per-inference latency with 0.70M parameters, yielding 3-10x lower latency than the Transformer baselines evaluated on the same hardware, while maintaining competitive accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2510_05255
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Rivaling Transformers: Multi-Scale Structured State-Space Mixtures for Agentic 6G O-RAN
Rezazadeh, Farhad
Chergui, Hatim
Debbah, Merouane
Song, Houbing
Niyato, Dusit
Liu, Lingjia
Networking and Internet Architecture
In sixth-generation (6G) Open Radio Access Networks (O-RAN), proactive control is preferable. A key open challenge is delivering control-grade predictions within Near-Real-Time (Near-RT) latency and computational constraints under multi-timescale dynamics. We therefore cast RAN Intelligent Controller (RIC) analytics as an agentic perceive-predict xApp that turns noisy, multivariate RAN telemetry into short-horizon per-User Equipment (UE) key performance indicator (KPI) forecasts to drive anticipatory control. In this regard, Transformers are powerful for sequence learning and time-series forecasting, but they are memory-intensive, which limits Near-RT RIC use. Therefore, we need models that maintain accuracy while reducing latency and data movement. To this end, we propose a lightweight Multi-Scale Structured State-Space Mixtures (MS3M) forecaster that mixes HiPPO-LegS kernels to capture multi-timescale radio dynamics. We develop stable discrete state-space models (SSMs) via bilinear (Tustin) discretization and apply their causal impulse responses as per-feature depthwise convolutions. Squeeze-and-Excitation gating dynamically reweights KPI channels as conditions change, and a compact gated channel-mixing layer models cross-feature nonlinearities without Transformer-level cost. The model is KPI-agnostic -- Reference Signal Received Power (RSRP) serves as a canonical use case -- and is trained on sliding windows to predict the immediate next step. Empirical evaluations conducted using our bespoke O-RAN testbed KPI time-series dataset (59,441 windows across 13 KPIs). Crucially for O-RAN constraints, MS3M achieves a 0.057 s per-inference latency with 0.70M parameters, yielding 3-10x lower latency than the Transformer baselines evaluated on the same hardware, while maintaining competitive accuracy.
title Rivaling Transformers: Multi-Scale Structured State-Space Mixtures for Agentic 6G O-RAN
topic Networking and Internet Architecture
url https://arxiv.org/abs/2510.05255