Výsledky vyhledávání - "Goodall, Alexander"

Report

Leveraging Approximate Model-based Shielding for Probabilistic Safety Guarantees in Continuous Environments

Autor: Goodall, Alexander W., Belardinelli, Francesco

Shielding is a popular technique for achieving safe reinforcement learning (RL). However, classical shielding approaches come with quite restrictive assumptions making them difficult to deploy in complex environments, particularly those with continuo

Externí odkaz: http://arxiv.org/abs/2402.00816

Zobrazit plný text záznamu

Report

Approximate Model-Based Shielding for Safe Reinforcement Learning

Autor: Goodall, Alexander W., Belardinelli, Francesco

Reinforcement learning (RL) has shown great potential for solving complex tasks in a variety of domains. However, applying RL to safety-critical systems in the real-world is not easy as many algorithms are sample-inefficient and maximising the standa

Externí odkaz: http://arxiv.org/abs/2308.00707

Zobrazit plný text záznamu

Report

Approximate Shielding of Atari Agents for Safe Exploration

Autor: Goodall, Alexander W., Belardinelli, Francesco

Balancing exploration and conservatism in the constrained setting is an important problem if we are to use reinforcement learning for meaningful tasks in the real world. In this paper, we propose a principled algorithm for safe exploration based on t

Externí odkaz: http://arxiv.org/abs/2304.11104

Zobrazit plný text záznamu

Report

Trieste: Efficiently Exploring The Depths of Black-box Functions with TensorFlow

Autor: Picheny, Victor, Berkeley, Joel, Moss, Henry B., Stojic, Hrvoje, Granta, Uri, Ober, Sebastian W., Artemev, Artem, Ghani, Khurram, Goodall, Alexander, Paleyes, Andrei, Vakili, Sattar, Pascual-Diaz, Sergio, Markou, Stratis, Qing, Jixiang, Loka, Nasrulloh R. B. S, Couckuyt, Ivo

We present Trieste, an open-source Python package for Bayesian optimization and active learning benefiting from the scalability and efficiency of TensorFlow. Our library enables the plug-and-play of popular TensorFlow-based models within sequential d

Externí odkaz: http://arxiv.org/abs/2302.08436

Zobrazit plný text záznamu