Multi-Label Indexing Technology for News with AI-Based Text Processing

Metadata

Publisher: SMPTE
Doc Type: Journal Article
Content Type: Original Research
Abbreviated Title: SMPTE Motion Imaging Jour.
Volume: 134, No. 6, pp. 16–24
Abstract: Broadcasting organizations produce large volumes of news articles daily, requiring accurate metadata to enable efficient reuse across television and online platforms. Manual annotation, however, is both time-consuming and laborintensive. To address this, we propose an AI-based system for automatic multilabel classification of news text. A central challenge in this task is the imbalanced label distribution, where high-frequency labels dominate and rare labels are underrepresented. To mitigate this, we introduce Weighted Asymmetric Loss (WASL) with Label Smoothing, which integrates class-balanced weighting, suppression of dominant negative samples, and smoothing based on label co-occurrence to improve performance on infrequent labels. Evaluation on Reuters-21578 and Japan Broadcasting Corp. (NHK) News Web datasets demonstrates that our approach significantly outperforms baseline methods on both macro-F1 and rnicro-F1 scores. We further developed a prototype system and deployed it in local NHK broadcasting stations, where it reduced metadata creation costs and facilitated content reuse, while high-lighting practical considerations for workflow adaptation.
Publication Date: 2025-10-01
DOI: 10.5594/JMI.2025/OJUX5368
ISSN: Print: 1545-0279 | Electronic: 2160-2492
Link: https://doi.org/10.5594/JMI.2025/OJUX5368
Author(s): Yuki Yasuda
Simon Clippingdale
Taro Miyazaki
Jun Goto
Takahiro Mochizuki
Keyword(s): Natural Language Processing, Multi-Label Text Classification, Loss Function, Few-Shot Learning
Copyright: © 2025 SMPTE

Source Data (JSON)

Full registry record with provenance metadata. Open directly: /api/doc/10.5594-JMI.2025-OJUX5368.json

Reference this Doc

Plain text (ISO 690 compliant)

Preview:

Yuki Yasuda, Simon Clippingdale, Taro Miyazaki, Jun Goto, and Takahiro Mochizuki; Multi-Label Indexing Technology for News with AI-Based Text Processing, MIJ 2025, Volume 134, Number 6 (pp. 16 to 24); SMPTE, 2025. Available at https://doi.org/10.5594/JMI.2025/OJUX5368

Snippet:

Yuki Yasuda, Simon Clippingdale, Taro Miyazaki, Jun Goto, and Takahiro Mochizuki; Multi-Label Indexing Technology for News with AI-Based Text Processing, MIJ 2025, Volume 134, Number 6 (pp. 16 to 24); SMPTE, 2025. Available at https://doi.org/10.5594/JMI.2025/OJUX5368

HTML (ISO 690 compliant)

Preview:

Snippet:

<span class="citation">Yuki Yasuda, Simon Clippingdale, Taro Miyazaki, Jun Goto, and Takahiro Mochizuki; <cite>Multi-Label Indexing Technology for News with AI-Based Text Processing</cite>, MIJ 2025, Volume 134, Number 6 (pp. 16 to 24); SMPTE, 2025. Available at <a href="https://doi.org/10.5594/JMI.2025/OJUX5368" target="_blank" rel="noopener">https://doi.org/10.5594/JMI.2025/OJUX5368</a></span>

SMPTE's HTML Pub

Preview:

Yuki Yasuda, Simon Clippingdale, Taro Miyazaki, Jun Goto, and Takahiro Mochizuki; Multi-Label Indexing Technology for News with AI-Based Text Processing, MIJ 2025, Volume 134, Number 6 (pp. 16 to 24); SMPTE, 2025
doi: 10.5594/JMI.2025/OJUX5368
url: https://doi.org/10.5594/JMI.2025/OJUX5368

Snippet:

<li>
Yuki Yasuda, Simon Clippingdale, Taro Miyazaki, Jun Goto, and Takahiro Mochizuki; <cite id="bib-10-5594-jmi-2025-ojux5368">Multi-Label Indexing Technology for News with AI-Based Text Processing</cite>, MIJ 2025, Volume 134, Number 6 (pp. 16 to 24); SMPTE, 2025
<span class="doi">10.5594/JMI.2025/OJUX5368</span>
</li>