no code implementations • 28 Mar 2024 • Yujin Chen, Yinyu Nie, Benjamin Ummenhofer, Reiner Birkl, Michael Paulitsch, Matthias Müller, Matthias Nießner
In Mesh2NeRF, we propose an analytic solution to directly obtain ground-truth radiance fields from 3D meshes, characterizing the density field with an occupancy function featuring a defined surface thickness, and determining view-dependent color through a reflection function considering both the mesh and environment lighting.
1 code implementation • 26 Mar 2024 • Kai Yuan, Christoph Bauinger, Xiangyi Zhang, Pascal Baehr, Matthias Kirchhart, Darius Dabert, Adrien Tousnakhoff, Pierre Boudier, Michael Paulitsch
We compare our approach to a similar CUDA implementation for MLPs and show that our implementation on the Intel Data Center GPU outperforms the CUDA implementation on Nvidia's H100 GPU by a factor up to 2. 84 in inference and 1. 75 in training.
no code implementations • 6 Nov 2023 • Gabriela Ben Melech Stan, Diana Wofk, Estelle Aflalo, Shao-Yen Tseng, Zhipeng Cai, Michael Paulitsch, Vasudev Lal
Our models are fine-tuned from existing pretrained models on datasets containing panoramic/high-resolution RGB images, depth maps and captions.
no code implementations • 31 Oct 2023 • Florian Geissler, Syed Qutub, Michael Paulitsch, Karthik Pattabiraman
We present a highly compact run-time monitoring approach for deep computer vision networks that extracts selected knowledge from only a few (down to merely two) hidden layers, yet can efficiently detect silent data corruption originating from both hardware memory and input faults.
1 code implementation • 30 Oct 2023 • Ralf Graafe, Qutub Syed Sha, Florian Geissler, Michael Paulitsch
Transient or permanent faults in hardware can render the output of Neural Networks (NN) incorrect without user-specific traces of the error, i. e. silent data errors (SDE).
no code implementations • 14 Sep 2023 • Syed Sha Qutub, Neslihan Kose, Rafael Rosales, Michael Paulitsch, Korbinian Hagn, Florian Geissler, Yang Peng, Gereon Hinz, Alois Knoll
The proposed loss functions in BEA improve the confidence score calibration and lower the uncertainty error, which results in a better distinction of true and false positives and, eventually, higher accuracy of the object detection models.
no code implementations • 15 Mar 2023 • Rafael Rosales, Pablo Munoz, Michael Paulitsch
We compare ensembles created with diverse model architectures trained either independently or through a Neural Architecture Search technique and evaluate the correlation of prediction-based and attribution-based diversity to the final ensemble accuracy.
no code implementations • 3 Mar 2023 • Rafael Rosales, Peter Popov, Michael Paulitsch
The key idea is to create successive model experts for samples that were difficult (not necessarily incorrectly classified) by the preceding model.
no code implementations • 9 Dec 2022 • Neslihan Kose, Ranganath Krishnan, Akash Dhamasia, Omesh Tickoo, Michael Paulitsch
Reliable uncertainty quantification in deep neural networks is very crucial in safety-critical applications such as automated driving for trustworthy and informed decision-making.
1 code implementation • 7 Sep 2022 • Syed Qutub, Florian Geissler, Yang Peng, Ralf Grafe, Michael Paulitsch, Gereon Hinz, Alois Knoll
The evaluation of several representative object detection models shows that even a single bit flip can lead to a severe silent data corruption event with potentially critical safety implications, with e. g., up to (much greater than) 100 FPs generated, or up to approx.
no code implementations • 16 Aug 2021 • Florian Geissler, Syed Qutub, Sayanta Roychowdhury, Ali Asgari, Yang Peng, Akash Dhamasia, Ralf Graefe, Karthik Pattabiraman, Michael Paulitsch
Convolutional neural networks (CNNs) have become an established part of numerous safety-critical computer vision applications, including human robot interactions and automated driving.
no code implementations • 30 Sep 2020 • Florian Geissler, Alex Unnervik, Michael Paulitsch
Trustworthy environment perception is the fundamental basis for the safe deployment of automated agents such as self-driving vehicles or intelligent robots.