Subhabrata Mukherjee

Subhabrata Mukherjee is a prominent researcher in the fields of Machine Learning and Natural Language Processing. His work focuses on advancing techniques such as model compression, multimodal learning, and credibility analysis. He holds a Ph.D. from Saarland University (2017), where he explored probabilistic graphical models for credibility analysis in online communities. His research spans topics including large language models (LLMs), knowledge distillation, and ethical AI.

Key contributions include the development of retrieval-augmented generation frameworks, cost-efficient LLM architectures, and systems for mitigating misinformation. His work bridges theoretical advancements with practical applications in healthcare, social media, and multilingual systems. Collaborations with institutions like Microsoft and academic networks highlight his interdisciplinary approach to AI challenges.

Research Highlights:

• Developed GraphRAG for integrating graph structures into generative models
• Pioneered Hybrid LLM architectures balancing efficiency and quality
• Advanced techniques for hallucination reduction in LLMs
• Contributed to fake news detection systems like DeClarE
• Explored experience-aware recommendation models

His publications reflect sustained innovation across NLP, ML systems, and ethical AI, with impactful contributions to both academic conferences (ACL, NeurIPS, KDD) and industry research.