|
Vrushank Ahire I'm a senior CS undergrad at the Indian Institute of Technology (IIT) Ropar, advised by Dr. Mudasir Ganaie. My research focuses on multimodal AI, with a particular emphasis on speech and vision, as well as ML optimization. I've gained research experience through internships at NTU-Singapore, Peking University, and Monash University. My work has been published in top conferences and journals, including CVPR, IJCNN, HiPC, Neural Networks and Pattern Recognition. I am also a recipient of the Hyundai Hope Scholarship. |
|
ResearchMy research spans low-resource and multilingual ASR, multimodal deepfake and emotion recognition, as well as reasoning in LLMs. I also work on clinical AI with EEG and neuroimaging data, along with formulating ML models. |
2026
Universum-enhanced GEPSVM classifiers improving EEG seizure detection through stable eigenvalue formulations.
2025
Bi-directional cross-modal attention across visual/audio/text modalities for dynamic emotion recognition in polar coordinates.
Combines granular ball computing with Twin SVM for robust multi-class classification solving two smaller QP problems for efficiency.
Ensemble fusing spatial transformers and frequency analysis with conditional face segmentation for state-of-the-art deepfake detection.
Graph-embedded intuitionistic fuzzy RVFL network tackling class imbalance while preserving dataset topology.
Integrates granular balls concept and Universum learning into Twin SVM for noise-robust classification with enhanced domain knowledge.
2024
Hybrid cache replacement algorithm combining random allocation with V-Way cache achieving up to 80.82% hit rates.
|
design adapted from Jon Barron's website |