A small collection of browser-based probability and stochastic-process demonstrations. These pages are meant to be played with: change parameters, simulate paths, and compare empirical behavior with the mathematical model.
Differential Privacy
A brief introduction to some of the main concepts of differential privacy.
Monte Carlo methods
These are the slides, in Turkish, that I presented in a couple of short courses about Monte Carlo methods.
Expectation-maximization
Here you will find a derivation that shows why the EM algorithm guarantees that the likelihood does not decrease over the iterations.
Hidden Markov Models, Forward filtering-Backward smoothing, Kalman Filter, Particle Filter
These are the slides that I presented at TESA: Laboratoire coopératif de recherche en Télécommunications Spatiales et Aéronautiques, Toulouse, France, on 12 June 2023. It is a brief introduction to Hidden Markov models (HMMs) and Bayesian inference in HMMs. You will find easy-to-implement pseducodes for forward filtering-backward smoothing for finite state-space HMMs, Kalman filtering (and smoothing) for linear Gaussian HMMs and particle filtering for general HMMs.
Yeni Fikirler Çalıştayı was held on 18-19 January 2025 in Mimar Sinan Güzel Sanatlar Üniversitesi, Bomonti Konferans Salonu. The event was very fulfilling for those who wanted to hear and share ideas in Machine Learning and Optimization. On behalf of the organizing committee (Figen Öztoprak, İlker Birbil, Kamer Kaya, Özgür Martin, Sinan Yıldırım), I thank everyone who attended, especially those who gave a talk.
The talks are available here.