pith. machine review for the scientific record. sign in

arxiv: 2604.07272 · v1 · submitted 2026-04-08 · 💻 cs.CL

Recognition: 2 theorem links

· Lean Theorem

ClickGuard: A Trustworthy Adaptive Fusion Framework for Clickbait Detection

Chhavi Dhiman , Naman Chawla , Riya Dhami , Gaurav Kumar , Ganesh Naik
Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:52 UTC · model grok-4.3

classification 💻 cs.CL
keywords clickbait detectionadaptive fusionBERT embeddingsCNN-BiLSTMinterpretabilityLIMEtrustworthy AIonline credibility
0
0 comments X

The pith

ClickGuard fuses BERT embeddings with structural features via an adaptive block and hybrid CNN-BiLSTM to reach 96.93 percent accuracy on clickbait detection.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents ClickGuard as a framework that combines BERT embeddings for semantic content with structural features through a Syntactic-Semantic Adaptive Fusion Block. This integrated representation is processed by a hybrid CNN-BiLSTM to capture both local patterns and sequential dependencies in headlines. The model reports 96.93 percent test accuracy while exceeding prior methods, and it applies LIME and permutation feature importance to assess which inputs drive predictions and how sensitive the output is to feature changes. Ablation experiments confirm that the fusion block contributes to the performance gains. The work positions the approach as a practical way to flag misleading headlines and support more credible online information.

Core claim

ClickGuard integrates BERT embeddings and structural features using the Syntactic-Semantic Adaptive Fusion Block for dynamic integration, followed by a hybrid CNN-BiLSTM to capture patterns and dependencies, achieving 96.93% testing accuracy while demonstrating interpretability and robustness through LIME and Permutation Feature Importance analysis.

What carries the argument

The Syntactic-Semantic Adaptive Fusion Block (SSAFB), which dynamically integrates BERT embeddings and structural features before they enter the hybrid CNN-BiLSTM classifier.

If this is right

  • The model outperforms state-of-the-art approaches on the tested datasets.
  • Ablation studies show that removing the adaptive fusion block reduces performance.
  • LIME and permutation importance analysis indicate the model is sensitive to changes in key syntactic and semantic features.
  • The framework scales across diverse datasets while maintaining high accuracy and providing interpretability.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same fusion strategy could be tested on related tasks such as fake news or sensationalist social media post detection.
  • Real-time filtering systems would need additional latency measurements to confirm practical deployment speed.
  • Retraining or fine-tuning on newer headline collections could be required as clickbait tactics evolve.

Load-bearing premise

The datasets used for training and evaluation represent the range of clickbait styles that appear in real-world online content.

What would settle it

Applying the trained model to a fresh collection of recent headlines from multiple news sources and measuring whether accuracy falls substantially below 90 percent would test the generalization claim.

Figures

Figures reproduced from arXiv: 2604.07272 by Chhavi Dhiman, Ganesh Naik, Gaurav Kumar, Naman Chawla, Riya Dhami.

Figure 1
Figure 1. Figure 1: Proposed ClickGuard Architecture; SFG: Sentence Feature Generation; POS: Part-of-Speech; SSAFB: Syntactic-Semantic Adaptive Fusion Block; MHA: Multihead Attention 𝐻௦௘௤ = {(𝑖𝑛𝑝𝑢𝑡_𝑖𝑑𝑠௜ , 𝑎𝑡𝑡𝑒𝑛𝑡𝑖𝑜𝑛_𝑚𝑎𝑠𝑘𝑠௜ , 𝑡𝑜𝑘𝑒𝑛_𝑡𝑦𝑝𝑒_𝑖𝑑𝑠௜)| 𝑖 = 1,2, . . , 𝑁} (3) where 𝐻௦௘௤ is the sequence of tokenized embeddings for all N samples. Next, the tokenized embeddings are processed using Multi-Head Attention (MHA) with four attenti… view at source ↗
Figure 2
Figure 2. Figure 2: Histogram of features for (a) Clickbait and (b) Non-Clickbait Headlines from Dataset1 [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Top 20 (𝑎) Clickbait and (𝑏) non-clickbait feature word clouds generated from Dataset1 ("make," "know") and adjectives ("real," "best") enhance the appeal, while abundant punctuation and slang terms (e.g., "bae," "dead") create a conversational and attention-grabbing style. In contrast, non-clickbait titles ( [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Syntactic-Semantic Adaptive Fusion Block (SSAFB) [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Radar charts depicting feature distribution [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Preprocessing Pipeline Architecture for SFG features [PITH_FULL_IMAGE:figures/full_fig_p015_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Training vs Validation accuracy for Dataset 1(a), Dataset 2(c), Dataset 3(e) and Training vs Validation loss for dataset 1(b), Dataset [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Perturbation Analysis (dataset1) The comparative evaluation of the proposed model against state-of-the-arts [5], [7], [10], [13], [23] and [25] as reported in [PITH_FULL_IMAGE:figures/full_fig_p019_9.png] view at source ↗
read the original abstract

The widespread use of clickbait headlines, crafted to mislead and maximize engagement, poses a significant challenge to online credibility. These headlines employ sensationalism, misleading claims, and vague language, underscoring the need for effective detection to ensure trustworthy digital content. The paper introduces, ClickGuard: a trustworthy adaptive fusion framework for clickbait detection. It combines BERT embeddings and structural features using a Syntactic-Semantic Adaptive Fusion Block (SSAFB) for dynamic integration. The framework incorporates a hybrid CNN-BiLSTM to capture patterns and dependencies. The model achieved 96.93% testing accuracy, outperforming state-of-the-art approaches. The model's trustworthiness is evaluated using LIME and Permutation Feature Importance (PFI) for interpretability and perturbation analysis. These methods assess the model's robustness and sensitivity to feature changes by measuring the average prediction variation. Ablation studies validated the SSAFB's effectiveness in optimizing feature fusion. The model demonstrated robust performance across diverse datasets, providing a scalable, reliable solution for enhancing online content credibility by addressing syntactic-semantic modelling challenges. Code of the work is available at: https://github.com/palindromeRice/ClickBait_Detection_Architecture

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper presents ClickGuard, a framework for clickbait detection that fuses BERT embeddings with structural features via a Syntactic-Semantic Adaptive Fusion Block (SSAFB) and a hybrid CNN-BiLSTM model. It reports a peak test accuracy of 96.93%, outperforming prior state-of-the-art methods, and evaluates trustworthiness through LIME, Permutation Feature Importance (PFI), perturbation analysis, and ablation studies demonstrating the value of the adaptive fusion component. The work claims robust performance across diverse datasets and provides open-source code.

Significance. If the performance and generalization claims hold under rigorous evaluation, the adaptive syntactic-semantic fusion mechanism could offer a practical, interpretable advance in clickbait detection, supporting more trustworthy online content moderation. The inclusion of interpretability tools and ablation results, together with public code, strengthens the contribution relative to purely empirical detection papers.

major comments (3)
  1. [Experimental Results] Experimental Results and Evaluation sections: The central claim of 96.93% test accuracy and outperformance of state-of-the-art methods is load-bearing for the paper's contribution, yet the manuscript supplies no dataset names, sizes, class balances, collection dates, train-test split protocol, or confirmation that test headlines are temporally or source-disjoint from training data. Without these details and accompanying statistical significance tests or error bars, the reported accuracy cannot be assessed for reproducibility or generalization.
  2. [Trustworthiness Evaluation] Trustworthiness and Robustness subsection: LIME, PFI, and perturbation analysis are conducted on the same data distribution used for training and testing. This does not address whether the SSAFB fusion remains stable under distribution shift (e.g., new headline sources or temporal drift), which directly undermines the claims of 'trustworthy' and 'scalable' performance across diverse real-world datasets.
  3. [Ablation Studies] Ablation Studies section: While ablation results are presented to validate the SSAFB, the paper does not report whether the fusion weights are learned end-to-end or fixed, nor does it quantify the contribution of each component relative to a simple concatenation baseline with the same CNN-BiLSTM backbone. This leaves open whether the reported gains are attributable to the adaptive mechanism or to other modeling choices.
minor comments (2)
  1. [Abstract] The abstract and introduction refer to 'diverse datasets' without naming them; adding explicit dataset citations and summary statistics in the main text would improve clarity.
  2. [Methodology] Notation for the SSAFB fusion weights and the hybrid CNN-BiLSTM layers should be defined consistently with equations or a diagram to avoid ambiguity in the architectural description.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. These have helped us identify areas where the manuscript can be strengthened for reproducibility, robustness evaluation, and clarity of contributions. We address each major comment below and indicate the revisions made.

read point-by-point responses

  1. Referee: [Experimental Results] Experimental Results and Evaluation sections: The central claim of 96.93% test accuracy and outperformance of state-of-the-art methods is load-bearing for the paper's contribution, yet the manuscript supplies no dataset names, sizes, class balances, collection dates, train-test split protocol, or confirmation that test headlines are temporally or source-disjoint from training data. Without these details and accompanying statistical significance tests or error bars, the reported accuracy cannot be assessed for reproducibility or generalization.

    Authors: We agree that these experimental details are necessary to support the central performance claims and enable reproducibility. In the revised manuscript, we have added a dedicated 'Datasets and Experimental Setup' subsection within the Experimental Results section. This includes the specific dataset names and sources, sizes, class balances, collection time periods, the exact train-validation-test split ratios and protocol, and explicit confirmation that test headlines are temporally and source-disjoint from training data. We have also added statistical significance testing (paired t-tests across multiple random seeds) and error bars to the performance tables. revision: yes

  2. Referee: [Trustworthiness Evaluation] Trustworthiness and Robustness subsection: LIME, PFI, and perturbation analysis are conducted on the same data distribution used for training and testing. This does not address whether the SSAFB fusion remains stable under distribution shift (e.g., new headline sources or temporal drift), which directly undermines the claims of 'trustworthy' and 'scalable' performance across diverse real-world datasets.

    Authors: The referee is correct that the original interpretability and perturbation analyses were performed on in-distribution data. While the perturbation analysis already quantifies sensitivity to feature changes, it does not fully substitute for explicit distribution-shift testing. In the revised version, we have added a new 'Robustness to Distribution Shift' paragraph in the Trustworthiness subsection. This includes results from a temporal hold-out experiment (training on earlier data and testing on later headlines) and reports the change in accuracy and fusion weight stability, thereby directly addressing stability under temporal drift. revision: yes

  3. Referee: [Ablation Studies] Ablation Studies section: While ablation results are presented to validate the SSAFB, the paper does not report whether the fusion weights are learned end-to-end or fixed, nor does it quantify the contribution of each component relative to a simple concatenation baseline with the same CNN-BiLSTM backbone. This leaves open whether the reported gains are attributable to the adaptive mechanism or to other modeling choices.

    Authors: We acknowledge the need for this clarification. The SSAFB fusion weights are learned end-to-end jointly with the rest of the model parameters. In the revised Ablation Studies section, we have explicitly stated this and added a new row in the ablation table comparing the full model against a non-adaptive baseline that performs simple concatenation of the BERT and structural features using the identical CNN-BiLSTM backbone. The results demonstrate that the adaptive fusion yields a statistically significant improvement over concatenation, isolating the contribution of the SSAFB. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical training and held-out evaluation on standard ML benchmarks

full rationale

The paper describes a neural architecture (BERT + structural features fused via SSAFB into CNN-BiLSTM) whose performance metric (96.93% test accuracy) is obtained by ordinary supervised training and evaluation on held-out splits. No equations, parameters, or self-citations are shown that would make the reported accuracy or interpretability scores (LIME/PFI) equivalent to the training inputs by construction. Ablation studies and SOTA comparisons are likewise experimental outcomes rather than definitional or self-referential reductions. The derivation chain is therefore self-contained against external data.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 1 invented entities

The central claim rests on standard deep-learning training assumptions plus the domain assumption that dynamic syntactic-semantic fusion improves detection over static concatenation.

free parameters (1)
  • Fusion weights and network hyperparameters
    Learned during training to optimize the adaptive fusion and classification performance on the given datasets.
axioms (2)
  • domain assumption BERT embeddings capture relevant semantic information for clickbait detection
    Invoked when using BERT as one input stream to the fusion block.
  • domain assumption LIME and permutation feature importance reliably indicate model trustworthiness
    Used to evaluate robustness and sensitivity without additional validation against human judgments.
invented entities (1)
  • Syntactic-Semantic Adaptive Fusion Block (SSAFB) no independent evidence
    purpose: Dynamically integrate BERT embeddings and structural features
    New component introduced by the paper to address syntactic-semantic modeling challenges.

pith-pipeline@v0.9.0 · 5521 in / 1231 out tokens · 33522 ms · 2026-05-10T17:52:25.334444+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

34 extracted references · 2 canonical work pages

  1. [1]

    Predicting Clickbait Strength in Online Social Media,

    V . Indurthi, B. Syed, M. Gupta and V . Varma, "Predicting Clickbait Strength in Online Social Media," in Proceedings of the 28th International Conference on Computational Linguistics, Barcelona, International Committee on Computational Linguistics, 2020, pp. 4835--4846

    2020

  2. [2]

    Detection and visualization of misleading content on Twitter,

    C. Boididou, S. Papadopoulos, . M. Zampoglou, L. Apostolidis, O. Papadopoulou and Y . Kompatsiaris, "Detection and visualization of misleading content on Twitter," International Journal of Multimedia Information Retrieval, vol. 7, pp. 71-86, 2018

    2018

  3. [3]

    Detecting Clickbait in Online Social Media: You Won’t Believe How We Did It,

    A. Elyashar, J. Bendahan and R. Puzis, "Detecting Clickbait in Online Social Media: You Won’t Believe How We Did It," in Cyber Security, Cryptology, and Machine Learning, Springer International Publishing, 2022, pp. 377-387

    2022

  4. [4]

    Experimental Evaluation of Clickbait Detection Using Machine Learning,

    I. Ahmad, M. A. Alqarni, A. A. Almazroi and A. Tariq, "Experimental Evaluation of Clickbait Detection Using Machine Learning," Intelligent Automation & Soft Computing, vol. 26, pp. 13335-1344, 2020

    2020

  5. [5]

    Prompt-tuning for Clickbait Detection via Text Summarization,

    H. Deng, Y . Zhu, Y . Wang, J. Qiang, Y . Yuan, Y . Li and . R. Zhang, "Prompt-tuning for Clickbait Detection via Text Summarization," arXiv:2404.11206, 2024

    work page arXiv 2024

  6. [6]

    Explainable hybrid word representations for sentiment analysis of financial news,

    S. Adhikari, . S. Thapa and U. Naser, "Explainable hybrid word representations for sentiment analysis of financial news," Neural Networks, vol. 164, pp. 115-123, 2023

    2023

  7. [7]

    Baitradar: A Multi-Model Clickbait Detection Algorithm Using Deep Learning,

    B. Gamage, A. Labib, A. Joomun, C. H. Lim and K. Wong, "Baitradar: A Multi-Model Clickbait Detection Algorithm Using Deep Learning," in ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021, pp. 2665-2669

    2021

  8. [8]

    A syntactic dependency method for aspect-level sentiment classification by deep learning,

    S. Chen, X. Du, J. Zhao, H. Huang and X. Chen, "A syntactic dependency method for aspect-level sentiment classification by deep learning," Measurement and Control, vol. 56, pp. 1057-1065, 2023

    2023

  9. [9]

    Blockchain- Enabled Deep Recurrent Neural Network Model for Clickbait Detection,

    A. Razaque, B. Alotaibi, M. Alotaibi, F. Amsaad, A. Manasov, S. Hariri and B. Yergaliyeva, "Blockchain- Enabled Deep Recurrent Neural Network Model for Clickbait Detection," IEEE Access, vol. 10, pp. 3144- 3163, 2022

    2022

  10. [10]

    An Attention-Based Neural Network Using Human Semantic Knowledge and Its Application to Clickbait Detection,

    F. Wei and U. T. Nguyen, "An Attention-Based Neural Network Using Human Semantic Knowledge and Its Application to Clickbait Detection," IEEE Open Journal of the Computer Society, vol. 3, pp. 217--232, 2022

    2022

  11. [11]

    Clickbait detection in telugu: Overcoming nlp challenges in resource-poor languages using benchmarked techniques,

    M. Marreddy, . S. R. Oota, L. S. Vakada, . V . C. Chinni and R. Mamidi, "Clickbait detection in telugu: Overcoming nlp challenges in resource-poor languages using benchmarked techniques," International Joint Conference on Neural Networks (IJCNN), pp. 1-8, 2021

    2021

  12. [12]

    Explainable sentiment analysis: a hierarchical transformer-based extractive summarization approach.,

    L. Bacco, A. Cimino, F. Dell’Orletta and M. Merone, "Explainable sentiment analysis: a hierarchical transformer-based extractive summarization approach.," Electronics, vol. 10, p. 2195, 2021

    2021

  13. [13]

    Detecting clickbaits using two-phase hybrid CNN-LSTM biterm model,

    S. Kaur, P. Kumar and P. Kumaraguru, "Detecting clickbaits using two-phase hybrid CNN-LSTM biterm model," Expert Systems with Applications, vol. 151, p. 113350, 2020

    2020

  14. [14]

    A polarization fusion network with geometric feature embedding for SAR ship classification,

    T. Zhang and X. Zhang, "A polarization fusion network with geometric feature embedding for SAR ship classification," Pattern Recognition, vol. 123, p. 108365, 2022

    2022

  15. [15]

    Clickbait Detection with Style-aware Title Modeling and Co-attention,

    C. Wu, F. Wu, T. Qi and Y . Huang, "Clickbait Detection with Style-aware Title Modeling and Co-attention," in Chinese Computational Linguistics: 19th China National Conference, CCL 2020, Hainan, China, October 30--November 1, 2020, Proceedings 19, Springer, 2020, pp. 430-443

    2020

  16. [16]

    A Deep Multi-level Attentive Network for Multimodal Sentiment Analysis,

    A. Yadav and D. K. Vishwakarma, "A Deep Multi-level Attentive Network for Multimodal Sentiment Analysis," ACM Transactions on Multimedia Computing, Communications, and Applications, vol. 19, no. 1, pp. 1-19, 2023

    2023

  17. [17]

    Modality compensation network: Cross-modal adaptation for action recognition,

    S. Song, J. Liu, Y . Li and Z. Guo, "Modality compensation network: Cross-modal adaptation for action recognition," IEEE Transactions on Image Processing, vol. 29, pp. 3957-3969, 2020

    2020

  18. [18]

    A deep learning framework for clickbait detection on social area network using natural language cues,

    B. Naeem, A. Khan, M. O. Beg and H. Mujtaba, "A deep learning framework for clickbait detection on social area network using natural language cues," Journal of Computational Social Science, vol. 3, 2020

    2020

  19. [19]

    Clickbait detection using multiple categorisation techniques,

    A. Pujahari and D. S. Sisodia, "Clickbait detection using multiple categorisation techniques," Journal of Information Science, vol. 47, pp. 118-128, 2021

    2021

  20. [20]

    Clickbait detection on WeChat: A deep model integrating semantic and syntactic information,,

    T. Liu, K. Yu, L. Wang, X. Zhang, H. Zhou and X. Wu, "Clickbait detection on WeChat: A deep model integrating semantic and syntactic information,," Knowledge-Based Systems, 2022

    2022

  21. [21]

    LSTM Powered Identification of Clickbait Content on Entertainment and News Websites,

    N. Bhoj, A. R. Dwivedi, A. Tripathi and B. Pandey, "LSTM Powered Identification of Clickbait Content on Entertainment and News Websites," in 2021 13th International Conference on Computational Intelligence and Communication Networks (CICN), vol. 2021, IEEE, 2021, pp. 29--33

    2021

  22. [22]

    Context-based Clickbait identification using Deep Learning,

    D. S. Thakur and S. Kurhade, "Context-based Clickbait identification using Deep Learning," in 2021 International Conference on Communication information and Computing Technology (ICCICT), Mumbai, 2021

    2021

  23. [23]

    Attention-fused deep relevancy matching network for clickbait detection,

    Q. Meng, B. Liu, X. Sun, H. Yan, C. Liang, J. Cao, R. K.-W. Lee and X. Bao, "Attention-fused deep relevancy matching network for clickbait detection," IEEE Transactions on Computational Social Systems, 2022

    2022

  24. [24]

    Intelligent Clickbait News Detection System Based on Artificial Intelligence and Feature Engineering,

    Y .-W. Ma, J.-L. Chen, L.-D. Chen and Y .-M. Huang, "Intelligent Clickbait News Detection System Based on Artificial Intelligence and Feature Engineering," IEEE Transactions on Engineering Management, pp. 1558-0040, 2022

    2022

  25. [25]

    Clickbait Detection via Contrastive Variational Modelling of Text and Label,

    X. Yi, J. Zhang, W. Li, X. Wang and X. Xie, "Clickbait Detection via Contrastive Variational Modelling of Text and Label," in Internation Joint Conference on Artificial Intelligence, Monteral, Canada, 2022

    2022

  26. [26]

    RGFN: Recurrent Graph Feature Network for ClickBait Detection,

    Y . Wang, H. Zhang, J. Zhu, Y . Li and L. Feng, "RGFN: Recurrent Graph Feature Network for ClickBait Detection," 2021 International Conference on High Performance Big Data and Intelligent Systems (HPBD&IS), pp. 151-156, 2021

    2021

  27. [27]

    MSynFD: Multi-hop Syntax aware Fake News Detection,

    L. Xiao, Q. Zhang, C. Shi, S. Wang, U. Naseem and L. Hu, "MSynFD: Multi-hop Syntax aware Fake News Detection," in Association for Computing Machinery, New York, 2024

    2024

  28. [28]

    Clicks can be Cheating: Counterfactual Recommendation for Mitigating Clickbait Issue,

    W. Wang, F. Feng, X. He, H. Zhang and T.-S. Chua, "Clicks can be Cheating: Counterfactual Recommendation for Mitigating Clickbait Issue," in Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, Canada, 2021

    2021

  29. [29]

    GAN-based Unsupervised Clickbait Style Transfer

    M. Agarwal and S. Kundu, "GAN-based Unsupervised Clickbait Style Transfer"

  30. [30]

    Stop Clickbait: Detecting and preventing clickbaits in online news media,

    A. Chakraborty, B. Paranjape, S. Kakarla and N. Ganguly, "Stop Clickbait: Detecting and preventing clickbaits in online news media," in 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), 2016, pp. 9-16

    2016

  31. [31]

    Clickbait Detection,

    S. R. Khater, O. H. Al-sahlee, D. M. Daoud and M. S. A. El-Seoud, "Clickbait Detection," in Proceedings of the 7th international conference on software and information engineering, 2018, pp. 111-115

    2018

  32. [32]

    Clickbait detection using deep learning,

    A. Agrawal, "Clickbait detection using deep learning," in 2016 2nd International Conference on Next Generation Computing Technologies (NGCT), 2016, pp. 268-272

    2016

  33. [33]

    Learning to Identify Ambiguous and Misleading News Headlines,

    W. Wei and X. Wan, "Learning to Identify Ambiguous and Misleading News Headlines," arXiv preprint arXiv:1705.06031, 2017

    work page arXiv 2017

  34. [34]

    Detecting incongruity between news headline and body text via a deep hierarchical encoder,

    S. Yoon, K. Park, J. Shin, H. Lim, S. Won, M. Cha and K. Jung, "Detecting incongruity between news headline and body text via a deep hierarchical encoder," in Proceedings of the AAAI conference on artificial intelligence, 2019, pp. 791-800

    2019