Zobrazeno 1 - 9
of 9
pro vyhledávání: '"Michael Truong Le"'
Publikováno v:
ICCV
We consider adversarial examples for image classification in the black-box decision-based setting. Here, an attacker cannot access confidence scores, but only the final label. Most attacks for this scenario are either unreliable or inefficient. Focus
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::93865d3bde7c1d4bc4d700b548ba3cc3
https://doi.org/10.1109/iccv.2019.00506
https://doi.org/10.1109/iccv.2019.00506
Autor:
Klemens Esterle, Patrick Hart, Daniel Malovetz, Michael Truong Le, Martin Buechel, Tobias Kessler, Alois Knoll, Thomas Brunner, Julian Bernhard, Frederik Diehl
Publikováno v:
2019 IEEE Intelligent Vehicles Symposium (IV).
Although many research vehicle platforms for autonomous driving have been built in the past, hardware design, source code and lessons learned have not been made available for the next generation of demonstrators. This raises the efforts for the resea
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::5bdeed98b59ce8339d640fa448d582e3
https://doi.org/10.1109/ivs.2019.8813784
https://doi.org/10.1109/ivs.2019.8813784
Publikováno v:
CVPR Workshops
Semantic Embeddings are a popular way to represent knowledge in the field of zero-shot learning. We observe their interpretability and discuss their potential utility in a safety-critical context. Concretely, we propose to use them to add introspecti
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::64eab772d715f8841a964337f8301416
http://arxiv.org/abs/1905.07733
http://arxiv.org/abs/1905.07733
By interpreting a traffic scene as a graph of interacting vehicles, we gain a flexible abstract representation which allows us to apply Graph Neural Network (GNN) models for traffic prediction. These naturally take interaction between traffic partici
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::4574651cc482846e0079f46a11f9c253
https://mediatum.ub.tum.de/1485457
https://mediatum.ub.tum.de/1485457
Graph Neural Networks (GNNs) research has concentrated on improving convolutional layers, with little attention paid to developing graph poolinglayers. Yet pooling layers can enable GNNs toreason over abstracted groups of nodes instead ofsingle nodes
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=od_______518::86e90f80b8a1eff4872c5b1bfe9cd8a9
https://mediatum.ub.tum.de/1521739
https://mediatum.ub.tum.de/1521739
Publikováno v:
ITSC
Object detection algorithms are essential components for perceiving the environment in safety-critical systems like automated driving. However, current state-of-the-art algorithms based on deep neural networks can give high confidence values to false
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::49c2854c8b95d22432165ca7001b9bef
https://doi.org/10.1109/itsc.2018.8569637
https://doi.org/10.1109/itsc.2018.8569637
Object detection algorithms are essential components for perceiving the environment in safety-critical systems like automated driving. However, current state-of-the-art algorithms based on deep neural networks can give high confidence values to false
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=od_______518::15ab1579450fb6867594a68163b79504
https://mediatum.ub.tum.de/1463390
https://mediatum.ub.tum.de/1463390
Publikováno v:
Intelligent Vehicles Symposium
In this paper, we compare different deep neural network approaches for motion prediction within a highway entrance scenario. The focus of our work lies on models that operate on limited history of data in order to fulfill the Markov property1 and be
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::1cee569374a7dab78387dc79f1653de3
https://doi.org/10.1109/ivs.2017.7995797
https://doi.org/10.1109/ivs.2017.7995797
Autor:
Harald Ruess, Yassine Hamza, Frederik Diehl, Georg Nuehrenberg, Michael Truong-Le, Chih-Hong Cheng, Markus Rickert, Gereon Hinz
Publikováno v:
DATE
We propose a methodology for designing dependable Artificial Neural Networks (ANN) by extending the concepts of understandability, correctness, and validity that are crucial ingredients in existing certification standards. We apply the concept in a c
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::1126982e4fcc839181ac94b858587219