# LEAF: Navigating Concept Drift in Cellular Networks

arXiv: https://arxiv.org/abs/2109.03011 · HTML: https://mlsanity.com/paper/2109.03011/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 87 references that are in the map.
- First posted on arXiv 7 Sep 2021.

## Builds on (most cited first, 15 of 15)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Random Erasing Data Augmentation](https://mlsanity.com/paper/1708.04896/) (2017)
- [Learning under Concept Drift: A Review](https://mlsanity.com/paper/2004.05785/) (2020)
- [On the Robustness of Interpretability Methods](https://mlsanity.com/paper/1806.08049/) (2018)
- [Visualizing the Effects of Predictor Variables in Black Box Supervised Learning Models](https://mlsanity.com/paper/1612.08468/) (2016)
- [CatBoost: gradient boosting with categorical features support](https://mlsanity.com/paper/1810.11363/) (2018)
- [On Wasserstein Two Sample Testing and Related Families of Nonparametric Tests](https://mlsanity.com/paper/1509.02237/) (2015)
- [Characterizing Concept Drift](https://mlsanity.com/paper/1511.03816/) (2015)
- [Correlation and variable importance in random forests](https://mlsanity.com/paper/1310.5726/) (2013)
- [Reactive Soft Prototype Computing for Concept Drift Streams](https://mlsanity.com/paper/2007.05432/) (2020)
- [Hot or not? Forecasting cellular network hot spots using sector performance indicators](https://mlsanity.com/paper/1704.05249/) (2017)
- [The Lockdown Effect: Implications of the COVID-19 Pandemic on Internet Traffic](https://mlsanity.com/paper/2008.10959/) (2020)
- [A Characterization of the COVID-19 Pandemic Impact on a Mobile Network Operator Traffic](https://mlsanity.com/paper/2010.02781/) (2020)
- [Characterizing Service Provider Response to the COVID-19 Pandemic in the United States](https://mlsanity.com/paper/2011.00419/) (2020)

## Built on by (most cited first, 9 of 9)

- [FlowWise: Stateful Fast-Slow Model Serving for Streaming Traffic Intelligence](https://mlsanity.com/paper/2402.03694/) (2024)
- [MYCROFT: Towards Effective and Efficient External Data Augmentation](https://mlsanity.com/paper/2410.08432/) (2024)
- [Algorithmic Data Minimization for Machine Learning over Internet-of-Things Data Streams](https://mlsanity.com/paper/2503.05675/) (2025)
- [A Representation Learning Approach to Feature Drift Detection in Wireless Networks](https://mlsanity.com/paper/2505.10325/) (2025)
- [Label-Free Concept Drift Assessment for Reliable AI in Emerging Wireless Applications](https://mlsanity.com/paper/2508.00042/) (2025)
- [TUBO: A Tailored ML Framework for Reliable Network Traffic Forecasting](https://mlsanity.com/paper/2602.11759/) (2026)
- [LLM-AUG: Robust Wireless Data Augmentation with In-Context Learning in Large Language Models](https://mlsanity.com/paper/2604.17770/) (2026)
- [ADORN: Adaptive Drift handling for Open RAN using Reinforcement Learning](https://mlsanity.com/paper/2607.08443/) (2026)
- [Maintaining IoT Device Identification under Concept Drift via Budget-Aware Traffic Labeling](https://mlsanity.com/paper/2608.15465/) (2026)

## Abstract (arXiv)

Operational networks commonly rely on machine learning models for many tasks, including detecting anomalies, inferring application performance, and forecasting demand. Yet, model accuracy can degrade due to concept drift, whereby the relationship between the features and the target to be predicted changes. Mitigating concept drift is an essential part of operationalizing machine learning models in general, but is of particular importance in networking's highly dynamic deployment environments. In this paper, we first characterize concept drift in a large cellular network for a major metropolitan area in the United States. We find that concept drift occurs across many important key performance indicators (KPIs), independently of the model, training set size, and time interval -- thus necessitating practical approaches to detect, explain, and mitigate it. We then show that frequent model retraining with newly available data is not sufficient to mitigate concept drift, and can even degrade model accuracy further. Finally, we develop a new methodology for concept drift mitigation, Local Error Approximation of Features (LEAF). LEAF works by detecting drift; explaining the features and time intervals that contribute the most to drift; and mitigates it using forgetting and over-sampling. We evaluate LEAF against industry-standard mitigation approaches (notably, periodic retraining) with more than four years of cellular KPI data. Our initial tests with a major cellular provider in the US show that LEAF consistently outperforms periodic and triggered retraining on complex, real-world data while reducing costly retraining operations.