Liquid Neural Network-based Adaptive Learning vs. Incremental Learning for Link Load Prediction amid Concept Drift due to Network Failures

TL;DR

This paper introduces a liquid neural network-based adaptive learning approach for traffic forecasting in communication networks, effectively handling sudden concept drifts caused by network failures without retraining.

Contribution

It presents a novel application of liquid neural networks for real-time adaptation to abrupt data distribution changes in network traffic prediction.

Findings

Liquid neural networks outperform incremental learning during drastic traffic pattern shifts.

The proposed method provides timely adaptation without retraining delays.

Experimental results demonstrate superior accuracy in failure scenarios.

Abstract

Adapting to concept drift is a challenging task in machine learning, which is usually tackled using incremental learning techniques that periodically re-fit a learning model leveraging newly available data. A primary limitation of these techniques is their reliance on substantial amounts of data for retraining. The necessity of acquiring fresh data introduces temporal delays prior to retraining, potentially rendering the models inaccurate if a sudden concept drift occurs in-between two consecutive retrainings. In communication networks, such issue emerges when performing traffic forecasting following a~failure event: post-failure re-routing may induce a drastic shift in distribution and pattern of traffic data, thus requiring a timely model adaptation. In this work, we address this challenge for the problem of traffic forecasting and propose an approach that exploits adaptive learning algorithms, namely, liquid neural networks, which are capable of self-adaptation to abrupt changes in data patterns without requiring any retraining. Through extensive simulations of failure scenarios, we compare the predictive performance of our proposed approach to that of a reference method based on incremental learning. Experimental results show that our proposed approach outperforms incremental learning-based methods in situations where the shifts in traffic patterns are drastic.