Sim-MSTNet: sim2real based Multi-task SpatioTemporal Network Traffic Forecasting

2026-01-29

AI Summary

This work addresses the challenges of weak generalization, task imbalance, and negative transfer in network traffic forecasting under data-scarce and multi-task settings. The authors propose Sim-MSTNet, the first approach to introduce the sim2real paradigm to this domain, leveraging synthetic data generation and domain randomization to construct diverse training samples. A bilevel optimization framework is designed to jointly learn sample weights and model parameters, while an attention mechanism combined with a dynamic loss weighting strategy enables collaborative multi-task optimization. This effectively mitigates negative transfer and enhances generalization under limited data. Experimental results on two open-source datasets demonstrate that the proposed method significantly outperforms existing approaches, achieving notable improvements in both prediction accuracy and generalization performance.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Network traffic forecasting plays a crucial role in intelligent network operations, but existing techniques often perform poorly when faced with limited data. Additionally, multi-task learning methods struggle with task imbalance and negative transfer, especially when modeling various service types. To overcome these challenges, we propose Sim-MSTNet, a multi-task spatiotemporal network traffic forecasting model based on the sim2real approach. Our method leverages a simulator to generate synthetic data, effectively addressing the issue of poor generalization caused by data scarcity. By employing a domain randomization technique, we reduce the distributional gap between synthetic and real data through bi-level optimization of both sample weighting and model training. Moreover, Sim-MSTNet incorporates attention-based mechanisms to selectively share knowledge between tasks and applies dynamic loss weighting to balance task objectives. Extensive experiments on two open-source datasets show that Sim-MSTNet consistently outperforms state-of-the-art baselines, achieving enhanced accuracy and generalization.

Problem

• network traffic forecasting
• data scarcity
• multi-task learning
• task imbalance
• negative transfer

Innovation

• sim2real
• multi-task learning
• domain randomization
• attention mechanism
• dynamic loss weighting