My research focuses on generative AI, diffusion models, trustworthy AI, and multimodal learning, with an emphasis on developing robust and uncertainty-aware methods for heterogeneous real-world data. I am currently a JST LOTUS Fellow at Osaka Metropolitan University, Japan.
I am a PhD Scholar in the Department of Computer Science and Engineering at the Indian Institute of Technology Roorkee, under the supervision of Prof. Balasubramanian Raman and Prof. Aparajita Khan. My research lies at the intersection of generative AI, diffusion models, multimodal learning, and trustworthy AI, with particular emphasis on uncertainty-aware and evidential deep learning for robust learning from noisy and heterogeneous data.
My broader research goal is to develop reliable multimodal learning systems that integrate complementary information across diverse modalities while providing calibrated and interpretable predictions. I apply these ideas to computer vision and scientific machine learning problems, including flood inundation mapping using multimodal satellite and multi-view imagery.
Prior to joining IIT Roorkee, I completed my M.Tech in Geoinformatics & Its Applications at Maulana Azad National Institute of Technology (MANIT), Bhopal, under the guidance of Prof. S. K. Katiyar, and my B.E. in Computer Science and Engineering from Government Engineering College, Raipur. I have been supported by the MHRD fellowship during my postgraduate and doctoral studies and am currently undertaking a research exchange at Osaka Metropolitan University through the JST LOTUS Fellowship.
Generative and diffusion-based architectures for robust representation learning, reconstruction, and prediction across heterogeneous data.
Learning complementary representations across heterogeneous modalities, including satellite, multi-view, and multi-sensor imagery.
Uncertainty quantification and evidential deep learning for calibrated, interpretable, and reliable predictions under uncertainty and noisy supervision.
Deep visual representation learning, segmentation, and information extraction across multiple viewpoints and sensing modalities.
AI-driven methods for scientific and environmental applications, with a focus on robust Earth observation and flood inundation mapping.
Indian Institute of Technology Roorkee · Advisors: Prof. Balasubramanian Raman, Prof. Aparajita Khan
Maulana Azad National Institute of Technology, Bhopal · Mentor: Prof. S. K. Katiyar
Government Engineering College, Raipur
I was selected for the JST LOTUS Fellowship exchange program and am currently a visiting researcher at Osaka Metropolitan University, Japan. My research collaboration focuses on trustworthy multimodal learning, uncertainty-aware AI, and generative learning methods.
During the fellowship, I am investigating diffusion-based generative learning and multimodal representation learning for reliable AI systems, extending my doctoral research toward broader computer vision and scientific machine learning applications.
