Modeling Romanized Hindi and Bengali: Dataset Creation and Multilingual LLM Integration

Keyphrases

• Language Model Integration 100%
• Multilingual Large Language Model 100%
• Transliteration 100%
• Bengali Dataset 100%
• Hindi Dataset 100%
• Dataset Generation 100%
• Indo-Aryan Languages 75%
• Romanized Script 75%
• Hindi 50%
• Multilingual Models 50%
• Bengali 50%
• Information Retrieval 25%
• Popular 25%
• Processing Task 25%
• Spoken Language 25%
• Speech Recognition 25%
• Sentiment Analysis 25%
• Natural Language Processing 25%
• Low-resource 25%
• Training Resources 25%
• Social Media Communication 25%
• Major Challenges 25%
• Still Face 25%
• Large Language Models 25%
• Seq2Seq 25%
• Digital Communication Platforms 25%
• Code-mixing 25%
• Intelligent Assistant 25%
• Resource Adaptation 25%
• Spelling Variation 25%
• South Asian Context 25%
• Large Language Models Training 25%
• Pronunciation Variation 25%
• Phonetic Structures 25%
• Roman alphabet 25%

Computer Science

• Large Language Model 100%
• Multilingual LLM 100%
• Experimental Result 50%
• Speech Recognition 50%
• Natural Language Processing 50%
• Sentiment Analysis 50%
• Spoken Language 50%
• Information Retrieval 50%
• Social Media 50%
• Personal Assistant 50%
• Roman Alphabet 50%
• Seq2Seq 50%