Multilingual clickbait detection with BERT: An empirical evaluation of training strategies

Keyphrases

• Training Strategy 100%
• Training Evaluation 100%
• Clickbait Detection 100%
• Cross-lingual Transfer 50%
• Multilingual BERT (mBERT) 50%
• Gemini 25%
• All Languages 25%
• Low-resource Settings 25%
• Robust Performance 25%
• Social Media Platforms 25%
• Practical Guidance 25%
• Second Language 25%
• Language Specificity 25%
• Error Analysis 25%
• Cosine 25%
• Transformer-based 25%
• Semantic Similarity 25%
• Urdu 25%
• Binary Label 25%
• BERT Model 25%
• Zero-shot 25%
• Practical Deployment 25%
• Unified Protocol 25%
• Multilingual Training 25%
• Bengali 25%
• Publicly Available Datasets 25%
• Multiple Languages 25%
• Multilingual Context 25%
• Web-based Media 25%
• Automated Content Analysis 25%
• Few-shot Prompting 25%
• Accuracy Comparison 25%
• Language Strategies 25%
• Error Aware 25%

Social Sciences

• Multilingualism 100%
• Bidirectional Encoder Representations From Transformers 100%
• Content Analysis 20%
• Foreign Languages 20%
• Individual Languages 20%
• Bengali 20%
• Zero-Shot Learning 20%
• Social Medium Platform 20%
• Semantic Similarity 20%

Computer Science

• Bidirectional Encoder Representations From Transformers 100%
• Content Analysis 25%
• False Negative 25%
• Cosine Similarity 25%
• Multiple Language 25%
• Language Specific 25%
• Analysis System 25%
• Unified Protocol 25%
• Zero-Shot Learning 25%
• Social Media 25%