Hi, I’m Khojasteh. I am a researcher at Mount Sinai NYC, Department of Neuroscience at the Schaffer Lab. I build artificial neural networks that help us understand how the brain works.
I am researching how a specific brain region (S1DZ) functions by building deep learning models with supervised and reinforcement learning. These models aim to mimic proprioceptive feedback. I then compare the network’s activations with real neural activity. I am also interested in the interpretability of networks, and understanding what’s happening inside them through reverse engineering.
My background is a mxiture of data science, ML, computer engineering, dynamcis and control, and some IoT. In my free time I enjoy watching Formula 1, flight simming, and sometimes tinkering with CAD/ 3D printing. I also like writing and making research more accessible, which is the purpose of this page. Science and research should be easy to understand for everyone. Avoiding complicated jargon provides a gentle introduction to the field. Distil (now discontinued) was an inspiration for this page. If you spot mistakes or have sugestions, feel free to reach out at kzm6[at]cornell[dot]edu. I am pretty active on LinkedIn as well.
Posts
-
Engineering a Tiny World Model: Representation, Latent Dynamics, and Rollout Stability
-
Implementing Different Types of RL Models
-
K-V Cache and Inference Optimization
-
Task-ID Conditioning as a Baseline for Multi-Task and Continual Learning
-
What is Feature Visualization in Neural Networks?
subscribe via RSS