288 citations · 556 across the 21 of their papers we have counts for
8 papers · 1 filter
CIMRL: Combining IMitation and Reinforcement Learning for Safe Autonomous Driving
Jonathan Booher, Khashayar Rohanimanesh, Junhong Xu +5
Modern approaches to autonomous driving rely heavily on learned components trained with large amounts of human driving data via imitation learning. However, these methods require l…
Uniformly Safe RL with Objective Suppression for Multi-Constraint Safety-Critical Applications
Zihan Zhou, Jonathan Booher, Khashayar Rohanimanesh +3
Safe reinforcement learning tasks are a challenging domain despite being very common in the real world. The widely adopted CMDP model constrains the risks in expectation, which mak…
Smoothed Embeddings for Certified Few-Shot Learning
Mikhail Pautov, Olesya Kuznetsova, Nurislam Tursynbek +2
Randomized smoothing is considered to be the state-of-the-art provable defense against adversarial perturbations. However, it heavily exploits the fact that classifiers map input o…
CC-Cert: A Probabilistic Approach to Certify General Robustness of Neural Networks
Mikhail Pautov, Nurislam Tursynbek, Marina Munkhoeva +3
In safety-critical machine learning applications, it is crucial to defend models against adversarial attacks -- small modifications of the input that change the predictions. Beside…
Manifold Hypothesis in Data Analysis: Double Geometrically-Probabilistic Approach to Manifold Dimension Estimation
Alexander Ivanov, Gleb Nosovskiy, Alexey Chekunov +7
Manifold hypothesis states that data points in high-dimensional space actually lie in close vicinity of a manifold of much lower dimension. In many cases this hypothesis was empiri…
Certified Robustness via Randomized Smoothing over Multiplicative Parameters of Input Transformations
Nikita Muravev, Aleksandr Petiushko
Currently the most popular method of providing robustness certificates is randomized smoothing where an input is smoothed via some probability distribution. We propose a novel appr…
Quadric Hypersurface Intersection for Manifold Learning in Feature Space
Fedor Pavutnitskiy, Sergei O. Ivanov, Evgeny Abramov +5
The knowledge that data lies close to a particular submanifold of the ambient Euclidean space may be useful in a number of ways. For instance, one may want to automatically mark an…
Robustness Threats of Differential Privacy
Nurislam Tursynbek, Aleksandr Petiushko, Ivan Oseledets
Differential privacy (DP) is a gold-standard concept of measuring and guaranteeing privacy in data analysis. It is well-known that the cost of adding DP to deep learning model is i…