Publications

SincPD, published open access in Discover Artificial Intelligence (Springer), 6(1), 2026. Adaptive Sinc-based bandpass filters learn informative frequency bands directly from raw vertical ground-reaction-force gait signals, reaching 98.77% accuracy for diagnosis and 97.22% for severity assessment. K-means clustering and silhouette scores prune the filter bank from 800 filters down to 30 with negligible loss, so the frequency bands driving each decision stay inspectable.

AdaptCMR is a parameter-efficient, detail-preserving deep learning framework for cardiac MRI reconstruction. It uses a spectrally guided mixture of experts to generalize across different views, contrasts, and acceleration factors, preserving fine anatomical detail with fewer parameters and faster inference.

Integrated FinBERT, a financial sentiment analysis LLM, to extract sentiment from news and combined it with stock prices in a PPO-LSTM model for stock trading. Achieved 134.39% cumulative return (2.34× initial budget) and a 1.46 Sharpe ratio, significantly outperforming DJIA, Ensemble, and PPO models.

Role-optimized agents (goalkeeping, defense, teamwork) in the 2D Simulation League; team placed 2nd at IranOpen 2023.