BERTO: an Adaptive BERT-based Network Time Series Predictor with Operator Preferences in Natural Language

TL;DR

BERTO addresses the need for accurate network time-series forecasting while offering operator-driven control over energy efficiency and service quality. It combines a BERT-based encoder with a time-series head and a prompt-driven Balancing Loss Function to steer predictions toward power-saving or QoS objectives via natural language prompts. Empirical results show a 4.13% improvement in MSE over strong baselines and demonstrate flexible trade-offs between energy savings and throughput loss across a realistic 428-cell deployment. This approach enables dynamic, parameter-free adaptation for intelligent RAN deployments, reducing manual tuning and enabling real-time objective alignment.

Abstract

We introduce BERTO, a BERT-based framework for traffic prediction and energy optimization in cellular networks. Built on transformer architectures, BERTO delivers high prediction accuracy, while its Balancing Loss Function and prompt-based customization allow operators to adjust the trade-off between power savings and performance. Natural language prompts guide the model to manage underprediction and overprediction in accordance with the operator's intent. Experiments on real-world datasets show that BERTO improves upon existing models with a $4.13$\% reduction in MSE while introducing the feature of balancing competing objectives of power saving and performance through simple natural language inputs, operating over a flexible range of $1.4$ kW in power and up to $9\times$ variation in service quality, making it well suited for intelligent RAN deployments.

Figures (4)

• Figure 1: System model.
• Figure 2: Solution Model.
• Figure 3: Prompt generation.
• Figure 4: BERTO Architecture.