Výsledky vyhledávání - "Jaunet, Theo"

Report

SIM2REALVIZ: Visualizing the Sim2Real Gap in Robot Ego-Pose Estimation

Autor: Jaunet, Theo, Bono, Guillaume, Vuillemot, Romain, Wolf, Christian

The Robotics community has started to heavily rely on increasingly realistic 3D simulators for large-scale training of robots on massive amounts of data. But once robots are deployed in the real world, the simulation gap, as well as changes in the re

Externí odkaz: http://arxiv.org/abs/2109.11801

Zobrazit plný text záznamu

Report

How Transferable are Reasoning Patterns in VQA?

Autor: Kervadec, Corentin, Jaunet, Theo, Antipov, Grigory, Baccouche, Moez, Vuillemot, Romain, Wolf, Christian

Since its inception, Visual Question Answering (VQA) is notoriously known as a task, where models are prone to exploit biases in datasets to find shortcuts instead of performing high-level reasoning. Classical methods address this by removing biases

Externí odkaz: http://arxiv.org/abs/2104.03656

Zobrazit plný text záznamu

Report

VisQA: X-raying Vision and Language Reasoning in Transformers

Autor: Jaunet, Theo, Kervadec, Corentin, Vuillemot, Romain, Antipov, Grigory, Baccouche, Moez, Wolf, Christian

Visual Question Answering systems target answering open-ended textual questions given input images. They are a testbed for learning high-level reasoning with a primary use in HCI, for instance assistance for the visually impaired. Recent research has

Externí odkaz: http://arxiv.org/abs/2104.00926

Zobrazit plný text záznamu

Report

DRLViz: Understanding Decisions and Memory in Deep Reinforcement Learning

Autor: Jaunet, Theo, Vuillemot, Romain, Wolf, Christian

We present DRLViz, a visual analytics interface to interpret the internal memory of an agent (e.g. a robot) trained using deep reinforcement learning. This memory is composed of large temporal vectors updated when the agent moves in an environment an

Externí odkaz: http://arxiv.org/abs/1909.02982

Zobrazit plný text záznamu

Akademický článek

Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.

Théo Guesser

Autor: Jaunet, Theo, Vuillemot, Romain, Wolf, Christian

Publikováno v: Workshop on Visualization for AI explainability (VISxAI)
Workshop on Visualization for AI explainability (VISxAI), Oct 2020, Salt-lake city, United States

International audience

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=dedup_wf_001::03527f50820dcae7b956b5f395edc25a
https://hal.archives-ouvertes.fr/hal-03347249

Zobrazit plný text záznamu

What if we Reduce the Memory of an Artificial Doom Player?

Autor: Jaunet, Theo, Vuillemot, Romain, Wolf, Christian

Publikováno v: Workshop on Visualization for AI explainability (VISxAI)
Workshop on Visualization for AI explainability (VISxAI), Oct 2019, Vancouver, Canada

International audience

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=dedup_wf_001::96f8448c6074701e22f6b35afdf342fc
https://hal.archives-ouvertes.fr/hal-03347241

Zobrazit plný text záznamu

RLMViz: Interpréter la Mémoire du Deep Reinforcement Learning

Autor: Jaunet, Theo, Vuillemot, Romain, Wolf, Christian

Publikováno v: Journée Visu 2019
Journée Visu 2019, May 2019, Paris, France

National audience; We present RLMViz, a visual analytics interface to interpret the internal memory of an agent (e.g., a robot) trained using deep reinforcement learning. This memory is composed of large temporal vectors updated before each action of

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=od______2592::2a9c1981889f7143580895ac9b712e56
https://hal.archives-ouvertes.fr/hal-02140902/file/Journee_Visu_2019__DRLViz.pdf

Zobrazit plný text záznamu