Research and Publications

Introduction

This paper presents Accelerated Weka, a GPU‑powered extension of the Weka Workbench that integrates high‑performance machine learning methods directly into Weka’s familiar graphical interface. By enabling GPU‑accelerated training while preserving Weka’s accessibility and ease of use, the framework dramatically reduces execution times on large datasets—achieving speedups of up to 2,198× on modern hardware—making advanced machine learning more approachable for students, practitioners, and researchers alike.

Problems addressed

• Traditional Weka implementations run on CPUs, which struggle with today’s large‑scale datasets.

• Many newcomers to machine learning find modern frameworks difficult due to their complex APIs and steep learning curves.

• GPU‑accelerated ML tools often require specialized knowledge, limiting accessibility for beginners.

• There is a need for a user‑friendly, open‑source environment that supports high‑performance ML without sacrificing simplicity.

Methods overview

Accelerated Weka integrates GPU‑optimized machine learning algorithms into the Weka Workbench while maintaining its intuitive GUI and workflow. The framework wraps GPU‑accelerated implementations behind Weka’s standard interface, enabling users to train models on GPUs without modifying their existing pipelines. Installation is streamlined through a Conda environment, and the system remains fully compatible with Weka’s GPL 3.0 licensing and plugin ecosystem.

Key outcomes

• Introduces GPU‑accelerated machine learning directly within the Weka Workbench.

• Achieves up to 2,198× speedup on large datasets using modern GPUs.

• Maintains Weka’s intuitive GUI, making high‑performance ML accessible to beginners.

• Provides a simple installation process via Conda.

• Preserves Weka’s open‑source GPL 3.0 license and plugin compatibility.

Experimental findings

Benchmark experiments show that Accelerated Weka delivers massive performance gains, with speedups reaching 2,198× on an NVIDIA A100 GPU. These improvements significantly reduce training time for large datasets while preserving Weka’s usability, demonstrating that GPU acceleration can be seamlessly integrated into an educational and beginner‑friendly ML environment.