Özgür Soysal

I cannot fiddle but I can make a great state from a little city (Themistocles, c.500 BC).

prof_pic.jpg

I am an EE PhD student at Stanford working on formalized ML and information theory.

During undergrad I worked on the mathematical foundations of machine learning advised by Orhan Arıkan and collaborated with Tolga Mete Duman to derive new bounds on the capacity of the binary deletion channel.

I also work on knowledge graph completion and goal planning for robotics in my free time.

If you are mathematically inclined and enjoy mini puzzles, please take a look at Jane Street, IBM, Bilkent

news

Sep 11, 2026 New bounds on the BDC capacity.
Oct 24, 2025 Presented the monomial paper at APSIPA 2025

selected publications

  1. bdc.png
    A New Upper Bound on the Binary Deletion Channel Capacity
    Ozgur Soysal
    Sep 2026
  2. monomial.png
    Monomial Matrix Relocation on the Loss Function Level-Set of Feedforward Neural Networks
    Ozgur Soysal, Arda Ozdemir, Yigit Yildirim, and 1 more author
    In 2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Sep 2025
  3. lowrank.png
    Low-Rank Compression of Neural Network Weights by Null-Space Encouragement
    Arda Ozdemir, Ozgur Soysal, Ege Doganay, and 2 more authors
    In 2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Sep 2025