I'm a Research Scientist at Google
Brain - Montreal.
Previously I was at DeepMind
I graduated in 2015 with a PhD in Computer Science from the Courant Institute. My advisor was Yann LeCun.
My interests (still) include: unsupervised learning, dictionary and representation (feature) learning, but recently also reinforcement learning and meta-learning.
In the past I have worked on computer vision and robotics projects which involved ideas from: 3D computer vision, machine learning, applied variational methods, geodesic active contours, and the level-set method.
Deep Learning Work
-Learning to Navigate in Complex Environments
Piotr Mirowski∗, Razvan Pascanu∗, Fabio Viola, Hubert Soyer, Andy Ballard, Andrea Banino, Misha Denil, Ross Goroshin, Laurent Sifre, Koray Kavukcuoglu, Dharshan Kumaran, Raia Hadsell
-Learning to Linearize Under Uncertainty (NIPS 2015)
Ross Goroshin, Michael Mathieu, Yann LeCun
-Unsupervised Learning of Spatiotemporally Coherent Metrics (ICCV 2015)
Ross Goroshin, Joan Bruna, Jonathan Tompson, David Eigen, Yann LeCun
-Unsupervised Feature Learning from Temporal Data (NIPS 2014 DL-Workshop & ICLR 2015 Workshop)
Ross Goroshin, Joan Bruna, Arthur Szlam, Jonathan Tompson, David Eigen, Yann LeCun
-Efficient Object Localization Using Convolutional Networks (CVPR 2015)
Jonathan Tompson, Ross Goroshin, Arjun Jain, Yann LeCun, Chris Bregler
Rostislav Goroshin and Yann LeCun, International Conference on Learning Representations (ICLR 2013)
-"Approximate solutions to several visibility optimization problems"
Rostislav Goroshin, Quyen Huynh, and Hao-Min Zhou, Communications in Mathematical Sciences Volume 9 Issue 2, June 2011
. Also available from UCLA(UCLA CAM Report 10-07
-"Automated cable detection in sonar imagery" J.C. Isaacs, R. Goroshin, IEEE SMC 2009
, San Antonio, TX
-"Obstacle Detection using a Monocular Camera" Rostislav Goroshin, Georgia Institute of Technology M.S. Thesis 2008. Georgia Tech ETD