A review of safe reinforcement learning: Methods, theory and applications

S Gu, L Yang, Y Du, G Chen, F Walter, J Wang… - arxiv preprint arxiv …, 2022 - arxiv.org
Reinforcement Learning (RL) has achieved tremendous success in many complex decision-
making tasks. However, safety concerns are raised during deploying RL in real-world …

A review of safe reinforcement learning: Methods, theories and applications

S Gu, L Yang, Y Du, G Chen, F Walter… - … on Pattern Analysis …, 2024 - ieeexplore.ieee.org
Reinforcement Learning (RL) has achieved tremendous success in many complex decision-
making tasks. However, safety concerns are raised during deploying RL in real-world …

Multi-task learning as a bargaining game

A Navon, A Shamsian, I Achituve, H Maron… - arxiv preprint arxiv …, 2022 - arxiv.org
In Multi-task learning (MTL), a joint model is trained to simultaneously make predictions for
several tasks. Joint training reduces computation costs and improves data efficiency; …

On the almost sure convergence of stochastic gradient descent in non-convex problems

P Mertikopoulos, N Hallak, A Kavis… - Advances in Neural …, 2020 - proceedings.neurips.cc
In this paper, we analyze the trajectories of stochastic gradient descent (SGD) with the aim of
understanding their convergence properties in non-convex problems. We first show that the …

AdaGrad avoids saddle points

K Antonakopoulos, P Mertikopoulos… - International …, 2022 - proceedings.mlr.press
Adaptive first-order methods in optimization have widespread ML applications due to their
ability to adapt to non-convex landscapes. However, their convergence guarantees are …

Gradient-descent quantum process tomography by learning Kraus operators

S Ahmed, F Quijandría, AF Kockum - Physical Review Letters, 2023 - APS
We perform quantum process tomography (QPT) for both discrete-and continuous-variable
quantum systems by learning a process representation using Kraus operators. The Kraus …

Riemannian stochastic optimization methods avoid strict saddle points

YP Hsieh, MR Karimi Jaghargh… - Advances in …, 2024 - proceedings.neurips.cc
Many modern machine learning applications-from online principal component analysis to
covariance matrix identification and dictionary learning-can be formulated as minimization …

Robust reinforcement learning via adversarial training with langevin dynamics

P Kamalaruban, YT Huang, YP Hsieh… - Advances in …, 2020 - proceedings.neurips.cc
We introduce a\emph {sampling} perspective to tackle the challenging task of training robust
Reinforcement Learning (RL) agents. Leveraging the powerful Stochastic Gradient Langevin …

Evaluating model-free reinforcement learning toward safety-critical tasks

L Zhang, Q Zhang, L Shen, B Yuan, X Wang… - Proceedings of the AAAI …, 2023 - ojs.aaai.org
Safety comes first in many real-world applications involving autonomous agents. Despite a
large number of reinforcement learning (RL) methods focusing on safety-critical tasks, there …

Mathematical introduction to deep learning: methods, implementations, and theory

A Jentzen, B Kuckuck, P von Wurstemberger - arxiv preprint arxiv …, 2023 - arxiv.org
This book aims to provide an introduction to the topic of deep learning algorithms. We review
essential components of deep learning algorithms in full mathematical detail including …