MambaCSP: Hybrid-Attention State Space Models for Hardware-Efficient Channel State Prediction

TL;DR

MambaCSP introduces a hybrid-attention state space model for efficient and accurate channel state prediction, outperforming LLM-based methods in speed and resource usage for wireless networks.

Contribution

The paper presents MambaCSP, a novel hybrid-attention SSM architecture that enhances CSI prediction efficiency and accuracy over traditional LLM-based models.

Findings

MambaCSP improves prediction accuracy by 9-12% over LLM-based approaches.

It achieves up to 3.0x higher throughput and 2.9x faster inference.

It reduces VRAM usage by 2.6x in simulations.

Abstract

Recent works have demonstrated that attention-based transformer and large language model (LLM) architectures can achieve strong channel state prediction (CSP) performance by capturing long-range temporal dependencies across channel state information (CSI) sequences. However, these models suffer from quadratic scaling in sequence length, leading to substantial computational cost, memory consumption, and inference latency, which limits their applicability in real-time and resource-constrained wireless deployments. In this paper, we investigate whether selective state space models (SSMs) can serve as a hardware-efficient alternative for CSI prediction. We propose MambaCSP, a hybrid-attention SSM architecture that replaces LLM-based prediction backbones with a linear-time Mamba model. To overcome the local-only dependencies of pure SSMs, we introduce lightweight patch-mixer attention layers that periodically inject cross-token attentions, helping with long-context CSI prediction. Extensive MISO-OFDM simulations show that MambaCSP improves prediction accuracy over LLM-based approaches by 9-12%, while delivering up to 3.0x higher throughput, 2.6x lower VRAM usage, and 2.9x faster inference. Our results demonstrate that hybrid state space architectures provide a promising direction for scalable and hardware-efficient AI-native CSI prediction in future wireless networks.