Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Analysis and Estimation of News Article Reading Time with Multimodal Machine Learning

Analysis and Estimation of News Article Reading Time with Multimodal Machine Learning

Analysis and Estimation of News Article Reading Time with Multimodal Machine Learning
Shotaro Ishihara, Yasufumi Nakama (IEEE BigData 2022, Industrial & Government Track)

Shotaro Ishihara

December 20, 2022
Tweet

More Decks by Shotaro Ishihara

Other Decks in Research

Transcript

  1. Shotaro Ishihara (Nikkei Inc.), and Yasufumi Nakama
    [email protected]
    IEEE BigData 2022, Industry and Government Program
    Does Text Length matter?
    Analysis and Estimation of News
    Article Reading Time with Multimodal
    Machine Learning

    View full-size slide

  2. Research Overview
    2
    ● text length
    ● headline / body text
    ● thumbnail image
    ● others like genre
    ● past reading history
    reading time

    View full-size slide

  3. Summary 1: Dataset
    3
    ● text length
    ● headline / body text
    ● thumbnail image
    ● others like genre
    ● past reading history
    reading time
    ✅ Real-world content and access log of Nikkei

    View full-size slide

  4. Summary 2: Text length
    4
    ● text length
    ● headline / body text
    ● thumbnail image
    ● others like genre
    ● past reading history
    reading time
    ✅ Doesn’t strongly correlate with reading time

    View full-size slide

  5. Summary 3: Multimodal
    5
    ● text length
    ● headline / body text
    ● thumbnail image
    ● others like genre
    ● past reading history
    reading time
    ✅ Boosted performance

    View full-size slide

  6. Outline
    6
    ● Introduction
    ● Problem Formulation
    ● Proposed Method
    ● Experiments
    ● Conclusion and Future Work

    View full-size slide

  7. Reading time estimation helps:
    7
    ● Push notifications [1]
    ● Recommendation [2, 4-6]
    ● User decision support [3, 7]
    ● Clickbait analysis [22-23]

    View full-size slide

  8. How can we estimate reading time?
    8
    ● text length
    ● headline / body text
    ● thumbnail image
    ● others like genre
    ● past reading history
    reading time

    View full-size slide

  9. Research questions
    9
    1. How much does text length correlate with
    reading time?
    2. How much do features other than text length
    improve the performance of reading time
    estimation?

    View full-size slide

  10. Reading time dataset
    10
    ● A large dataset that includes reading time from
    Japanese financial news from the Nikkei.
    ○ About 1,000 articles a day, 800,000 paid
    subscribers (and data infrastructure)
    ○ Larger and more scalable than some existing
    data on recording eye movements [8] [9] and
    brain activity [10]

    View full-size slide

  11. Dataset details
    11
    100,000 sessions * 3
    ● train: 21-12-01
    ● val: 21-12-08
    ● test: 21-12-15

    View full-size slide

  12. RQ1: text length (x) & reading time (y)
    12
    Correlation coefficient is 0.04 (and 0.31)

    View full-size slide

  13. 13
    ● Architecture
    corresponding to
    the specific data
    ● E2E fine-tuning
    Proposed Method

    View full-size slide

  14. Experiments: Features & Models
    14
    1. The model was fixed to LightGBM [16] and the
    features were explored.
    2. We fixed the features and observed differences.
    a. Ridge regression
    b. MLP
    c. Proposed method (w/wo E2E fine-tuning)

    View full-size slide

  15. Experiments: Features
    15
    Additional features improved the metric.

    View full-size slide

  16. 1. mean reading time
    2. text length
    3. minimum reading time
    4. embedding of body text (dimension 193)
    5. embedding of thumbnail image (dimension 88)
    Important features by LightGBM
    16

    View full-size slide

  17. Experiments: Models
    17
    ● LightGBM worked better in the same feature.
    ● Proposed method outperformed LightGBM by
    adding LSTM, and e2e fine-tuning.

    View full-size slide

  18. Multimodal training tips
    18
    ● Different learning rate: 2e-5 for BERT, 1e-4 for
    Swin Transformer, and 1e-2 for the others
    ● CosineAnnealingLR: For training stability

    View full-size slide

  19. Conclusion
    19
    ● We highlighted the importance of reading time
    and evaluated the implementation.
    ● Our analysis revealed reading time does not
    strongly correlate with text length.
    ● Our experiments showed a multimodal machine
    learning approach led to a more accurate
    estimation than simply using text length.

    View full-size slide

  20. Future Work
    20
    ● Offline evaluation => Online operation
    ● Further feature & model exploration
    ● Clickbait analysis

    View full-size slide