Výsledky vyhledávání - "Van Delft, Bastien"

Report

CODE: Confident Ordinary Differential Editing

Autor: van Delft, Bastien, Martorella, Tommaso, Alahi, Alexandre

Conditioning image generation facilitates seamless editing and the creation of photorealistic images. However, conditioning on noisy or Out-of-Distribution (OoD) images poses significant challenges, particularly in balancing fidelity to the input and

Externí odkaz: http://arxiv.org/abs/2408.12418

Zobrazit plný text záznamu

Report

Estimating Risk and Uncertainty in Deep Reinforcement Learning

Autor: Clements, William R., Van Delft, Bastien, Robaglia, Benoît-Marie, Slaoui, Reda Bahi, Toth, Sébastien

Reinforcement learning agents are faced with two types of uncertainty. Epistemic uncertainty stems from limited data and is useful for exploration, whereas aleatoric uncertainty arises from stochastic environments and must be accounted for in risk-se

Externí odkaz: http://arxiv.org/abs/1905.09638

Zobrazit plný text záznamu

TTT++: When Does Self-Supervised Test-Time Training Fail or Thrive?

Autor: Liu, Yuejiang, Kothari, Parth, Van Delft, Bastien, Bellot-Gurlet, Baptiste, Mordan, Taylor, Alahi, Alexandre

Test-time training (TTT) through self-supervised learning (SSL) is an emerging paradigm to tackle distributional shifts. Despite encouraging results, it remains unclear when this approach thrives or fails. In this work, we first provide an in-depth l

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=od_______185::05df4a077e810b1bc3ab00228de0386d
https://infoscience.epfl.ch/record/296742

Zobrazit plný text záznamu

Vyhledávací nástroje:

Upřesnit hledání