Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Transformers4NewsRec: A Transformer-based News Recommendation Framework

Dairui Liu, Honghui Du, Boming Yang, Neil Hurley, Aonghus Lawlor, Irene Li, Derek Greene, Ruihai Dong

TL;DR

Transformers4NewsRec, a new Python framework built on the \textbf{Transformers} library, is designed to unify and compare the performance of various news recommendation models, including deep neural networks and graph-based models.

Abstract

Pre-trained transformer models have shown great promise in various natural language processing tasks, including personalized news recommendations. To harness the power of these models, we introduce Transformers4NewsRec, a new Python framework built on the \textbf{Transformers} library. This framework is designed to unify and compare the performance of various news recommendation models, including deep neural networks and graph-based models. Transformers4NewsRec offers flexibility in terms of model selection, data preprocessing, and evaluation, allowing both quantitative and qualitative analysis.

Transformers4NewsRec: A Transformer-based News Recommendation Framework

TL;DR

Transformers4NewsRec, a new Python framework built on the \textbf{Transformers} library, is designed to unify and compare the performance of various news recommendation models, including deep neural networks and graph-based models.

Abstract

Pre-trained transformer models have shown great promise in various natural language processing tasks, including personalized news recommendations. To harness the power of these models, we introduce Transformers4NewsRec, a new Python framework built on the \textbf{Transformers} library. This framework is designed to unify and compare the performance of various news recommendation models, including deep neural networks and graph-based models. Transformers4NewsRec offers flexibility in terms of model selection, data preprocessing, and evaluation, allowing both quantitative and qualitative analysis.

Paper Structure

This paper contains 10 sections, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Framework: Transformers4NewsRec
  3. Data Module
  4. Model Module
  5. Evaluation and Utility Tools Module
  6. Evaluation of Transformers4NewsRec
  7. Model Performance Evaluation
  8. Efficiency Comparison of Batching Methods
  9. Topic-Centric Explanation Example
  10. Conclusion

Figures (2)

  • Figure 1: The overall architecture of the Transformers4NewsRec framework, illustrating its modular design. The framework integrates key components for data processing, model building, evaluation, and utility tools, enabling flexible experimentation and customization for news recommendation tasks.
  • Figure 2: Comparison between zero-padding and concatenation methods for data batching. The left side illustrates the traditional zero-padding method, while the right side shows the proposed concatenation method, which improves efficiency by reducing redundant padding during batch creation, ultimately speeding up training and evaluation.