AiGAS-dEVL-RC: An Adaptive Growing Neural Gas Model for Recurrently Drifting Unsupervised Data Streams

2025-04-08

AI Summary

To address the dual challenges of periodic concept drift and extreme label delay in unsupervised data streams, this paper proposes an adaptive Growing Neural Gas (GNG) model. The method integrates incremental clustering, concept memory compression, and a delayed-supervision response mechanism, establishing the first online learning framework that supports historical concept memory reuse and dynamic topological regeneration. It enables joint modeling and efficient reactivation of both recurring abrupt and gradual concept drifts. Evaluated on multiple concept drift benchmarks, the proposed approach achieves 12–23% higher prediction accuracy, reduces memory footprint by 40%, and shortens drift response latency to only 1–3 time steps—compared to state-of-the-art methods. These improvements significantly enhance practicality and robustness in resource-constrained and high-label-delay scenarios.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Natural Language Processing: Learning & Optimization for NLP
• Data Mining & Knowledge Management: Data Stream Mining

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Models for Web evolution

Abstract

Concept drift and extreme verification latency pose significant challenges in data stream learning, particularly when dealing with recurring concept changes in dynamic environments. This work introduces a novel method based on the Growing Neural Gas (GNG) algorithm, designed to effectively handle abrupt recurrent drifts while adapting to incrementally evolving data distributions (incremental drifts). Leveraging the self-organizing and topological adaptability of GNG, the proposed approach maintains a compact yet informative memory structure, allowing it to efficiently store and retrieve knowledge of past or recurring concepts, even under conditions of delayed or sparse stream supervision. Our experiments highlight the superiority of our approach over existing data stream learning methods designed to cope with incremental non-stationarities and verification latency, demonstrating its ability to quickly adapt to new drifts, robustly manage recurring patterns, and maintain high predictive accuracy with a minimal memory footprint. Unlike other techniques that fail to leverage recurring knowledge, our proposed approach is proven to be a robust and efficient online learning solution for unsupervised drifting data flows.

Problem

• Handles recurrent concept drifts in unsupervised data streams
• Adapts to incremental and abrupt distribution changes
• Manages extreme verification latency with minimal memory

Innovation

• Adaptive Growing Neural Gas for drifting data
• Handles abrupt and incremental concept drifts
• Efficient memory for recurring patterns