News Recommender System Considering Temporal Dynamics and Accuracy-Diversity Tradeoff
Bibliographic record
Abstract
News recommender systems aim to personalize users experience for online news readers and help them discover relevant and interesting news items from a broad and diverse search space. However, recommending news is a challenging task. There are hundreds of news articles published every day, many of which quickly become obsolete and irrelevant to the readers. Readers’ preferences (interests) also exhibit dynamic behavior and the relevance of readers’ preferences strongly depend on the context. Some of the readers’ preferences are long-term, reflecting the personality or behavior, whereas others are short-term, showing their current interests. External events, such as breaking news and trends, also influence readers’ interests. Although the high recommendation accuracy is appreciated, focusing on it too much sacrifices diversity and limits readers’ options. The main contribution of this research is to design a recommender system to tackle the specific challenges of news recommendations. To address these challenges, we propose a recommendation strategy that seamlessly integrates readers’ long-term and short-term preferences when recommending news items. In addition to high accuracy, we focus on promoting reasonable diversity in news recommendations. To achieve a balanced objective (high accuracy and reasonable diversity), we formulate a combined optimization strategy that includes both of these aspects in the recommendation process. More specifically, to achieve high accuracy, we propose novel latent factor models such as matrix factorization and a generalized linear model. In addition, we propose a regularized latent factor model to achieve reasonable diversity while maintaining high accuracy. Then, for the same purpose, we propose a deep learning-based framework composed of network components of news modelling and reader modelling. We use different regularization terms in the latent factor model to achieve accuracy-diversity balance, and the neural attention mechanism in deep neural networks for the same purpose. We also introduce a new evaluation metric to measure the tradeoff performance with respect to accuracy and diversity. Experiments on real-world data have demonstrated the effectiveness of our proposed approach on quality factors such as accuracy, diversity as well as this tradeoff metric. Our model has achieved an improvement on this new tradeoff metric by 10-30% when compared to traditional and state-of-the-art recommendation algorithms.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.003 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".