EDCA - An Evolutionary Data-Centric AutoML Framework for Efficient Pipelines

TL;DR

EDCA tackles the data-centric gap in AutoML by integrating data quality improvements and automatic data optimization into an evolutionary AutoML framework. It jointly optimizes data preprocessing, data reduction (instance/feature selection), and model configuration to form complete pipelines, evaluated against state-of-the-art systems on AMLB benchmarks. The results show EDCA achieves similar predictive performance to FLAML and TPOT while using substantially less data, illustrating a cost- and energy-efficient approach aligned with Green AutoML. This work highlights the practical impact of data-centric optimization for scalable, environmentally conscious ML pipelines and outlines future directions for broader benchmarking and parameter investigations.

Abstract

Automated Machine Learning (AutoML) gained popularity due to the increased demand for Machine Learning (ML) specialists, allowing them to apply ML techniques effortlessly and quickly. AutoML implementations use optimisation methods to identify the most effective ML solution for a given dataset, aiming to improve one or more predefined metrics. However, most implementations focus on model selection and hyperparameter tuning. Despite being an important factor in obtaining high-performance ML systems, data quality is usually an overlooked part of AutoML and continues to be a manual and time-consuming task. This work presents EDCA, an Evolutionary Data Centric AutoML framework. In addition to the traditional tasks such as selecting the best models and hyperparameters, EDCA enhances the given data by optimising data processing tasks such as data reduction and cleaning according to the problems' needs. All these steps create an ML pipeline that is optimised by an evolutionary algorithm. To assess its effectiveness, EDCA was compared to FLAML and TPOT, two frameworks at the top of the AutoML benchmarks. The frameworks were evaluated in the same conditions using datasets from AMLB classification benchmarks. EDCA achieved statistically similar results in performance to FLAML and TPOT but used significantly less data to train the final solutions. Moreover, EDCA experimental results reveal that a good performance can be achieved using less data and efficient ML algorithm aspects that align with Green AutoML guidelines

Figures (6)

• Figure 1: Process used by EDCA.
• Figure 2: Example of crossover. It is not applied to the IS genes because they are absent in both individuals. For the FS genes, a gene-level one-point crossover is applied. The non-DR genes use a uniform crossover at the parent level.
• Figure 3: Flow chart of the mutation applied to genes (IS or FS).
• Figure 4: Distribution of EDCA's DR techniques in a 5-fold CV across 30 runs.
• Figure 5: Number of pipelines evaluated by each framework over 30 runs.