Skip to content
aranyaadheu

RESEARCH

From questions to evidence.

My interests centre on NLP, large language models, and evaluating what works in practice. I’m developing these interests as an undergraduate research assistant at DeepNet Lab.

Conference paper

Impact of Data Augmentation Techniques on Bangla Cyberbullying Detection Using Transformer Models

Aniruddha Aranya A Dheu, Sayed Al Mamun, Sourav Mohon Biswas

3rd Student Research Conference (LURSSRC 2026) · Machine Learning & AI Track

Presented · August 2026

An empirical study of text data augmentation for transformer-based cyberbullying detection in Bangla, comparing models using standard classification metrics in a low-resource language setting.

  • Bangla NLP
  • Cyberbullying detection
  • Transformers
  • Data augmentation

Research interests

Language in context

Bengali–English code-mixed text, sentiment analysis, and the challenges of multilingual NLP.

Efficient adaptation

Parameter-efficient fine-tuning and how adaptation choices affect model performance and practical cost.

Grounded applications

Retrieval-augmented generation, document question answering, and useful language-model applications.

I also collect learning resources and research notes in my AI/ML Research Cookbook.

Research cookbook