Yaniv Nikankin

pic.jpg

I’m a CS PhD student at the Technion, where I’m advised by Prof. Yonatan Belinkov. My research interests revolve around Deep Learning and interpretability. I aim to study the properties and internals of neural networks, and utilize these understandings to build better models.

I completed my M.Sc under the supervision of Prof. Michal Irani, and prior to that I’ve worked as a software engineer in industry and received my B.Sc in Compuer science.

Publications

2024

  1. arithmetic_boh.png
    Arithmetic Without Algorithms: Language Models Solve Math With a Bag of Heuristics
    Yaniv Nikankin, Anja Reusch, Aaron Mueller, and Yonatan Belinkov
    2024

2023

  1. deconstructing.png
    Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses
    Gon Buzaglo, Niv Haim, Gilad Yehudai, Gal Vardi, Yakir Oz, Yaniv Nikankin, and Michal Irani
    In NeurIPS, 2023
  2. sinfusion.gif
    SinFusion: Training Diffusion Models on a Single Image or Video
    Yaniv Nikankin, Niv Haim, and Michal Irani
    In ICML, 2023