Ilker Yildirim

I am a research scientist at MIT. I mainly work with Josh Tenenbaum (MIT) and Winrich Freiwald (The Rockefeller University). Before, I did my Ph.D. studies at University of Rochester advised by Robbie Jacobs.
E: ilkery at
P: 585 2670718



My research takes a computational approach to perception and its neural basis. My scientific goal is to understand how individual sensory modalities (e.g., vision) and the interactions between more than one sensory modality (e.g., interactions between vision and touch) give rise to our seamless and richly structured everyday perceptual experiences. To this end, I utilize probabilistic programming and deep learning alongside psychophysics and computational neuroscience.

Recent talks
  • Human-like physical scene understanding by integrating a physics engine with deep learning, Physical and Social Scene Understanding Workshop at CogSci 16, Philadelphia, August, 2016.
  • Efficient analysis-by-synthesis in primate face processing, RIKEN Brain Science Institute, Tokyo, November 2015.
  • Efficient analysis-by-synthesis in primate face processing, International Symposium on Object Vision in Humans, Monkeys, and Machines at the University of Electro-Communications, Tokyo, November 2015.
Recent Abstracts and Workshop papers

Data Set

Tutorials and Code

Below are short notes on Bayesian inference and Bayesian nonparametrics. I provided sample code alongside each note hoping that it will be helpful. Feel free to drop me a line about what you think, and if you see any bugs. See the Computational Cognition Cheat sheets for an extensive list of tutorials, including the ones below.