Zobrazeno 1 - 10
of 31
pro vyhledávání: '"Koehler, Gregor"'
Numerical weather prediction (NWP) models struggle to skillfully predict tropical precipitation occurrence and amount, calling for alternative approaches. For instance, it has been shown that fairly simple, purely data-driven logistic regression mode
Externí odkaz:
http://arxiv.org/abs/2401.03746
Autor:
Kades, Klaus, Sellner, Jan, Koehler, Gregor, Full, Peter M, Lai, T Y Emmy, Kleesiek, Jens, Maier-Hein, Klaus H
Publikováno v:
JMIR Medical Informatics, Vol 9, Iss 2, p e22795 (2021)
BackgroundNatural Language Understanding enables automatic extraction of relevant information from clinical text data, which are acquired every day in hospitals. In 2018, the language model Bidirectional Encoder Representations from Transformers (BER
Externí odkaz:
https://doaj.org/article/1669c6e2993142dc9b667062979d37cd
Regularization is a critical component in deep learning training, with weight decay being a commonly used approach. It applies a constant penalty coefficient uniformly across all parameters. This may be unnecessarily restrictive for some parameters,
Externí odkaz:
http://arxiv.org/abs/2311.09058
Autor:
Koehler, Gregor, Wald, Tassilo, Ulrich, Constantin, Zimmerer, David, Jaeger, Paul F., Franke, Jörg K. H., Kohl, Simon, Isensee, Fabian, Maier-Hein, Klaus H.
Despite the remarkable success of deep learning systems over the last decade, a key difference still remains between neural network and human decision-making: As humans, we cannot only form a decision on the spot, but also ponder, revisiting an initi
Externí odkaz:
http://arxiv.org/abs/2309.07513
Autor:
Ma, Jun, Zhang, Yao, Gu, Song, Ge, Cheng, Ma, Shihao, Young, Adamo, Zhu, Cheng, Meng, Kangkang, Yang, Xin, Huang, Ziyan, Zhang, Fan, Liu, Wentao, Pan, YuanKe, Huang, Shoujin, Wang, Jiacheng, Sun, Mingze, Xu, Weixin, Jia, Dengqiang, Choi, Jae Won, Alves, Natália, de Wilde, Bram, Koehler, Gregor, Wu, Yajun, Wiesenfarth, Manuel, Zhu, Qiongjie, Dong, Guoqiang, He, Jian, Consortium, the FLARE Challenge, Wang, Bo
Quantitative organ assessment is an essential step in automated abdominal disease diagnosis and treatment planning. Artificial intelligence (AI) has shown great potential to automatize this process. However, most existing AI algorithms rely on many e
Externí odkaz:
http://arxiv.org/abs/2308.05862
Autor:
Wald, Tassilo, Ulrich, Constantin, Isensee, Fabian, Zimmerer, David, Koehler, Gregor, Baumgartner, Michael, Maier-Hein, Klaus H.
Independently trained machine learning models tend to learn similar features. Given an ensemble of independently trained models, this results in correlated predictions and common failure modes. Previous attempts focusing on decorrelation of output pr
Externí odkaz:
http://arxiv.org/abs/2307.02516
Autor:
Roy, Saikat, Wald, Tassilo, Koehler, Gregor, Rokuss, Maximilian R., Disch, Nico, Holzschuh, Julius, Zimmerer, David, Maier-Hein, Klaus H.
Foundation models have taken over natural language processing and image generation domains due to the flexibility of prompting. With the recent introduction of the Segment Anything Model (SAM), this prompt-driven paradigm has entered image segmentati
Externí odkaz:
http://arxiv.org/abs/2304.05396
Autor:
Roy, Saikat, Koehler, Gregor, Baumgartner, Michael, Ulrich, Constantin, Petersen, Jens, Isensee, Fabian, Maier-Hein, Klaus
Owing to success in the data-rich domain of natural images, Transformers have recently become popular in medical image segmentation. However, the pairing of Transformers with convolutional blocks in varying architectural permutations leaves their rel
Externí odkaz:
http://arxiv.org/abs/2304.04225
Autor:
Roy, Saikat, Koehler, Gregor, Ulrich, Constantin, Baumgartner, Michael, Petersen, Jens, Isensee, Fabian, Jaeger, Paul F., Maier-Hein, Klaus
There has been exploding interest in embracing Transformer-based architectures for medical image segmentation. However, the lack of large-scale annotated medical datasets make achieving performances equivalent to those in natural images challenging.
Externí odkaz:
http://arxiv.org/abs/2303.09975
Autor:
Lüth, Carsten T., Zimmerer, David, Koehler, Gregor, Jaeger, Paul F., Isensee, Fabian, Petersen, Jens, Maier-Hein, Klaus H.
Unsupervised anomaly detection in medical imaging aims to detect and localize arbitrary anomalies without requiring annotated anomalous data during training. Often, this is achieved by learning a data distribution of normal samples and detecting anom
Externí odkaz:
http://arxiv.org/abs/2301.02126