Jekyll2026-06-04T15:45:53+00:00https://kiarashkianid.github.io/feed.xmlKiarash KianidehkordiData Science, Computational Cognition, and AI Research PortfolioKiarash Kianidehkordikiarash.kianidehkordi@mail.utoronto.caPredict Human Reading Time Using GPT-2 & BERT2026-05-04T00:00:00+00:002026-05-04T00:00:00+00:00https://kiarashkianid.github.io/reading-time-gpt2-bertCourse project for Seminar in Computational Cognition.
Preprocessed the Natural Stories Dataset with more than 1M observations and engineered features for reading time prediction.
Built a sliding-window batching pipeline to split text into segments for parallel, CUDA-efficient inference of GPT-2 and BERT models.
Analyzed surprisal scores and fitted linear mixed-effects models to predict human reading time.
Contributed quantitative evidence on GPT-2 outperforming BERT with an AIC improvement of 571 points and p-value less than 1e-100.
]]>Kiarash Kianidehkordikiarash.kianidehkordi@mail.utoronto.caParkinson Classification & Symptoms Profiling with Accelerometer Data2026-05-03T00:00:00+00:002026-05-03T00:00:00+00:00https://kiarashkianid.github.io/parkinson-accelerometerInter-University Health Data and AI Inquiry Program project.
Applied Gaussian mixture model clustering to 400+ participant wrist-worn accelerometer data to classify Parkinson’s disease versus controls.
Achieved a Silhouette Score of 0.688 without labeled training data.
Engineered tremor-specific features via bandpass filtering between 3 and 12 Hz and PSD metrics.
Produced symptom profiles with direct applications in remote health monitoring.