CLeAN: Continual Learning Adaptive Normalization in Dynamic Environments

Isabella Marasco; Davide Evangelista; Elena Loli Piccolomini; Michele Colajanni

cs.LGMar 18, 2026

CLeAN: Continual Learning Adaptive Normalization in Dynamic Environments

Authors: Isabella Marasco, Davide Evangelista, Elena Loli Piccolomini, Michele Colajanni

Organizations: Department of Computer Science and Engineering University of Bologna Bologna, 40126, Italy

Abstract

Artificial intelligence systems predominantly rely on static data distributions, making them ineffective in dynamic real-world environments, such as cybersecurity, autonomous transportation, or finance, where data shifts frequently. Continual learning offers a potential solution by enabling models to learn from sequential data while retaining prior knowledge. However, a critical and underexplored issue in this domain is data normalization. Conventional normalization methods, such as min-max scaling, presuppose access to the entire dataset, which is incongruent with the sequential nature of continual learning. In this paper we introduce Continual Learning Adaptive Normalization (CLeAN), a novel adaptive normalization technique designed for continual learning in tabular data. CLeAN involves the estimation of global feature scales using learnable parameters that are updated via an Exponential Moving Average (EMA) module, enabling the model to adapt to evolving data distributions. Through comprehensive evaluations on two datasets and various continual learning strategies, including Resevoir Experience Replay, A-GEM, and EwC we demonstrate that CLeAN not only improves model performance on new data but also mitigates catastrophic forgetting. The findings underscore the importance of adaptive normalization in enhancing the stability and effectiveness of tabular data, offering a novel perspective on the use of normalization to preserve knowledge in dynamic learning environments.

Figures & tables

Figure 1: A graphical representation of our proposed normalization scheme: CLeAN.

\displaystyle\begin{split}\hat{\boldsymbol{M}}_{t}&\leftarrow(1-\eta)\boldsymbol{M}_{t}+\eta\hat{\boldsymbol{M}}_{t-1}\\ \hat{\boldsymbol{m}}_{t}&\leftarrow(1-\eta)\boldsymbol{m}_{t}+\eta\hat{\boldsymbol{m}}_{t-1}\end{split}

Algorithm 1 The proposed CLeAN scheme

Figure 2: Performance comparison on the UNSW-NB15 dataset. (a–d) Average Accuracy, (e–h) Average AUROC, and (i–l) Average Forgetting are shown across the five daily experiences for different buffer strategies: (a, e, i) Finetuning, (b, f, j) Reservoir, (c, g, k) A-GEM, and (d, h, l) EwC. Each plot compares the results obtained with different normalization techniques (Global, Local, CN, CLeAN).

Figure 3: Performance comparison on the CICIDS-2017 dataset. (a-d) Average Accuracy, (e-h) Average AUROC, and (h-l) Average Forgetting are shown across the five daily experiences for different buffer strategies: (a, e, i) Finetuning, (b, f, j) Reservoir, (c, g, k) A-GEM, and (d, h, l) EwC. Each plot compares the results obtained with different normalization techniques (Global, Local, CN, CLeAN).

Continual Learning

Continual Learning studies how models acquire knowledge from changing data streams while retaining or selectively revising what they learned before.

See all 304 papers

TaskFusion: Continual Anomaly Detection for Heterogeneous Tabular Data
Jun 10, 2026Dayananda Herurkar, Federico Raue, Joachim Folz +2Continual Anomaly Detection Tabular Anomaly Detection
German Research Center for Artificial Intelligence (DFKI), Kaiserslautern, Germany · RPTU Kaiserslautern-Landau, Germany · Hochschule Bonn-Rhein-Sieg (H-BRS), Germany
Representation Finetuning for Continual Learning
Mar 11, 2026Haihua Luo, Xuming Ran, Tommi Kärkkäinen +4Continual Learning Fine-Tuning
Faculty of Information Technology, University of Jyväskylä, Jyväskylä, Finland · National University of Singapore, Singapore, Singapore · School of Computer Science and Technology, Dalian University of Technology, Dalian, China
NeuMoSync: End-to-End Neuromodulatory Control for Plasticity and Adaptability in Continual Learning
Aug 5, 2026Seyed Roozbeh Razavi Rohani, Khashayar Khajavi, Wesley Chung +2Class-Incremental Learning Dynamic Neural Networks
Department of Computer Science Simon Fraser University · Department of Computer Science McGill Univeristy, Mila - Quebec AI Institute

Jun 10, 2026J/K move · Enter open · S save

TaskFusion: Continual Anomaly Detection for Heterogeneous Tabular Data

Dayananda Herurkar, Federico Raue, Joachim Folz, Jörn Hees+1

German Research Center for Artificial Intelligence (DFKI), Kaiserslautern, Germany · RPTU Kaiserslautern-Landau, Germany · Hochschule Bonn-Rhein-Sieg (H-BRS), Germany

Continual Anomaly Detection Tabular Anomaly Detection Continual Learning OOD Detection Anomaly Detection

Open paper AskPDF arXiv

Continual anomaly detection in tabular data is challenging and remains largely underexplored, particularly in settings with heterogeneous feature schemas, distribution shifts, and severe class imbalance. In many real-world applications, data arrive sequentially from diverse domains, rendering conventional continual learning methods ineffective due to their reliance on a fixed input space. We propose a continual learning (CL) method, which can overcome these challenges and continually learn from different tasks. Our method consists of three main parts: our AGF model, Taskfusion augmentation, and outlier exposure. The AGF-model maps task-specific features into a shared space, then aligns distributions to reduce representation drift, and learns anomaly decision boundaries in the aligned space. To improve stability, we introduce Taskfusion augmentation, combining boundary-aware interpolation within tasks to refine the model anomaly boundaries and cross-task mixing to transfer anomaly structure across datasets. To handle class imbalance and memory constraints, we employ tabular dataset distillation to store compact synthetic replay samples, which are jointly used with augmented data in an outlier exposure objective for robust anomaly detection. We evaluate the approach on 21 heterogeneous datasets across multiple domains. Results show that our approach substantially improves continual anomaly detection performance over sequential fine-tuning and other CL baselines while reducing catastrophic forgetting and maintaining stable detection across heterogeneous datasets.

Abstract

Continual Learning

Explore similar work

TaskFusion: Continual Anomaly Detection for Heterogeneous Tabular Data

Representation Finetuning for Continual Learning

NeuMoSync: End-to-End Neuromodulatory Control for Plasticity and Adaptability in Continual Learning