profile photo

Hi! Iā€™m Nikita Durasov, a PhD student at EPFL IC (šŸ‡ØšŸ‡­) where I work under supervision of Professor Pascal Fua at Computer Vision Laboratory on problems connected with Deep Learning (šŸ¤–) and Computer Vision (šŸ‘).

Previously, I was a research scientist at Samsung AI and worked under supervision of Anton Konushin at Visual Understanding Lab. I did my bachelors at MIPT (šŸ‡·šŸ‡ŗ) with a degree in Applied Math jointly studying at Yandex School of Data Analysis.


Research


I'm broadly interested in improving interpretability and robustness of modern deep learning models through uncertainty quantification and bayesian methods.

As a result I'm also curious about new approaches to active learning, bayesian optimization and other data-efficient techiques for different domains.

Mostly, my research is focused on computer vision and image processing tasks but isn't limited to them. Occasionally, I'm trying to wrap my head around natural language processing and reinforcement learning.


Publications

depth estimation photo
Leveraging Self-Supervision for Cross-Domain Crowd Counting

Weizhe Liu, Nikita Durasov, Pascal Fua

arXiv 2021

pdf | code | video

depth estimation photo
Masksembles for Uncertainty Estimation

Nikita Durasov, Timur Bagautdinov, Pierre Baque, Pascal Fua

CVPR 2021

pdf | code | video | page

depth estimation photo
Double refinement network for efficient indoor monocular depth estimation

Nikita Durasov, Mikhail Romanov, Valeriya Bubnova, Pavel Bogomolov, Anton Konushin

IROS 2019

pdf | code | video

depth estimation photo
Learning node embeddings for influence set completion

Sergei Ivanov, Nikita Durasov, Evgeny Burnaev

ICDM 2018

pdf | code