Nikolay Savinov

PhD Student/Researcher

ETH Zurich/Google Brain

About

Welcome to my webpage! I’m a PhD student at ETH Zurich and a Student Researcher at Google Brain. My current research interests focus on memory for reinforcement learning, in particular, episodic memory. I’m also interested in applications of RL to robotics.

Interests

Deep Learning
Memory
Reinforcement Learning
Robotics
Computer Vision

Education

MSc in Applied Mathematics and Physics, 2013

Moscow Institute of Physics and Technology
BSc in Applied Mathematics and Physics, 2011

Moscow Institute of Physics and Technology

Experience

January 2019 – Present

Zurich

Student Researcher

Google Brain

Research areas: Memory, Exploration, Reinforcement Learning

Supervisors: Sylvain Gelly (Google Brain), Timothy Lillicrap (DeepMind)

June 2018 – December 2018

Zurich

Research Intern

Google Brain

Research areas: Memory, Exploration, Reinforcement Learning

Supervisors: Sylvain Gelly (Google Brain), Timothy Lillicrap (DeepMind)

June 2017 – December 2017

Munich

Research Intern

Intel Intelligent Systems Lab

Research areas: Memory, Navigation, Reinforcement Learning

Supervisors: Vladlen Koltun, Alexey Dosovitskiy

September 2013 – Present

Zurich

PhD Student/Research Assistant

ETH Zurich

Research areas: Computer Vision, Geometry, Deep Learning

Advisor: Prof. Marc Pollefeys

September 2011 – September 2013

Moscow

Software Engineer

Yandex

Machine Learning

Recent Publications

Episodic Curiosity through Reachability

Accepted to ICLR 2019

Nikolay Savinov, Anton Raichuk, Raphaël Marinier, Damien Vincent, Marc Pollefeys, Timothy Lillicrap, Sylvain Gelly

PDF Code Project Video

Conditional Affordance Learning for Driving in Urban Environments

Published at CoRL 2018 (oral)

Axel Sauer, Nikolay Savinov, Andreas Geiger

PDF Code Video

Semi-parametric Topological Memory for Navigation

Published at ICLR 2018

Nikolay Savinov, Alexey Dosovitskiy, Vladlen Koltun

PDF Code Project Video

Matching neural paths: transfer from recognition to correspondence search

Published at NIPS 2017

Nikolay Savinov, Lubor Ladicky, Marc Pollefeys

PDF Code Video

Quad-networks: unsupervised learning to rank for interest point detection

Published at CVPR 2017

Nikolay Savinov, Akihito Seki, Lubor Ladicky, Torsten Sattler, Marc Pollefeys

PDF Video

Semantic3D.net: A new Large-scale Point Cloud Classification Benchmark

Timo Hackel, Nikolay Savinov, Lubor Ladicky, Jan Wegner, Konrad Schindler, Marc Pollefeys

PDF Code Dataset

TI-POOLING: transformation-invariant pooling for feature learning in Convolutional Neural Networks

Published at CVPR 2016

Dmitry Laptev, Nikolay Savinov, Joachim Buhmann, Marc Pollefeys

PDF Code

Semantic 3D Reconstruction with Continuous Regularization and Ray Potentials Using a Visibility Consistency Constraint

Published at CVPR 2016 (spotlight oral)

Nikolay Savinov, Christian Haene, Lubor Ladicky, Marc Pollefeys

PDF Video

Discrete Optimization of Ray Potentials for Semantic 3D Reconstruction

Published at CVPR 2015

Nikolay Savinov, Lubor Ladicky, Christian Haene, Marc Pollefeys

PDF Video

Class Specific 3D Object Shape Priors Using Surface Normals

Published at CVPR 2014

Christian Haene, Nikolay Savinov, Marc Pollefeys

PDF

Recent Posts

Curiosity and Procrastination in Reinforcement Learning

Our post in Google AI Blog

Nikolay Savinov , Timothy Lillicrap

Last updated on Jan 10, 2019 2 min read