NewsReX: A More Efficient Approach to News Recommendation with Keras 3 and JAX

2025-08-29

AI Summary

News recommendation research faces three major challenges: poor reproducibility, high computational resource demands, and fragmented software ecosystems. To address these, we propose an efficient, open-source library built on Keras 3 and JAX, enabling cross-backend high-performance training optimization—including on low-resource GPUs (e.g., RTX 3060 Ti with 8 GB VRAM). Methodologically, the library unifies negative sampling strategies, batch processing mechanisms, and training configuration interfaces, establishing a standardized, reproducible experimental framework. It further supports rapid deployment of custom datasets and state-of-the-art news recommendation models. Experiments on the real-world Nikkei News dataset demonstrate substantial speedups in training time over mainstream implementations, without compromising model accuracy. This work advances news recommendation research toward reproducibility, lightweight deployment, and standardization.

Technology Category

• Data Mining & Knowledge Management: Recommender Systems
• Natural Language Processing: Learning & Optimization for NLP
• Machine Learning: Hardware-aware ML

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Reproducing and comparing results in news recommendation research has become increasingly difficult. This is due to a fragmented ecosystem of diverse codebases, varied configurations, and mainly due to resource-intensive models. We introduce NewsReX, an open-source library designed to streamline this process. Our key contribution is a modern implementation built on Keras 3 and JAX, which provides an increase in computational efficiency. Experiments show that NewsReX is faster than current implementations. To support broader research, we provide a straightforward guide and scripts for training models on custom datasets. We validated this functionality using a proprietary Japanese news dataset from Nikkei News, a leading Japanese media corporation renowned for its comprehensive coverage of business, economic, and financial news. NewsReX makes reproducing complex experiments faster and more accessible to a wider range of hardware making sure the speed up it also achieved for less powerful GPUs, like an 8GB RTX 3060 Ti. Beyond the library, this paper offers an analysis of key training parameters often overlooked in the literature, including the effect of different negative sampling strategies, the varying number of epochs, the impact of random batching, and more. This supplementary analysis serves as a valuable reference for future research, aiming to reduce redundant computation when comparing baselines and guide best practices. Code available at https://github.com/igor17400/NewsReX.

Problem

• Addresses fragmented ecosystem in news recommendation research
• Improves computational efficiency using Keras 3 and JAX
• Enables reproducible experiments across diverse hardware configurations

Innovation

• NewsReX library built on Keras 3
• JAX backend for computational efficiency
• Optimized for various hardware including GPUs