006 Spezielle Computerverfahren
Refine
Has Fulltext
- yes (11)
Is part of the Bibliography
- yes (11)
Document Type
- Journal article (11) (remove)
Keywords
- virtual reality (3)
- 3D model generation (1)
- Deep learning (1)
- Entscheidungsfindung (1)
- Ethik (1)
- Klima (1)
- Künstliche Intelligenz (1)
- Modell (1)
- Neuronales Netz (1)
- Optical Music Recognition (1)
- adaptive tutoring (1)
- artificial intelligence (1)
- authoring platform (1)
- avatar embodiment (1)
- avatars (1)
- background knowledge (1)
- body awareness (1)
- body image disturbance (1)
- body weight modification (1)
- body weight perception (1)
- classification (1)
- climate (1)
- complications (1)
- document analysis (1)
- eating and body weight disorders (1)
- embodiment (1)
- ethics (1)
- explainable AI (1)
- explanation complexity (1)
- fully convolutional neural networks (1)
- gambling (1)
- healthcare (1)
- healthcare professionals (1)
- higher education (1)
- historical document analysis (1)
- historical printings (1)
- human-centered AI (1)
- human-robot interaction (1)
- immersion (1)
- individual differences (1)
- machine learning (1)
- medieval manuscripts (1)
- model output statistics (1)
- neume notation (1)
- neural networks (1)
- object reconstruction (1)
- optical character recognition (1)
- oro-antral communication (1)
- oro-antral fistula (1)
- prediction (1)
- recommender agent (1)
- rehabilitation (1)
- risks (1)
- robot-supported training (1)
- robotic tutor (1)
- site mapping (1)
- structured light illumination (1)
- technology acceptance (1)
- technology-supported education (1)
- teeth extraction (1)
- therapy (1)
- underwater 3D scanning (1)
- user experience (1)
- virtual environments (1)
Institute
Sonstige beteiligte Institutionen
Climate models are the tool of choice for scientists researching climate change. Like all models they suffer from errors, particularly systematic and location-specific representation errors. One way to reduce these errors is model output statistics (MOS) where the model output is fitted to observational data with machine learning. In this work, we assess the use of convolutional Deep Learning climate MOS approaches and present the ConvMOS architecture which is specifically designed based on the observation that there are systematic and location-specific errors in the precipitation estimates of climate models. We apply ConvMOS models to the simulated precipitation of the regional climate model REMO, showing that a combination of per-location model parameters for reducing location-specific errors and global model parameters for reducing systematic errors is indeed beneficial for MOS performance. We find that ConvMOS models can reduce errors considerably and perform significantly better than three commonly used MOS approaches and plain ResNet and U-Net models in most cases. Our results show that non-linear MOS models underestimate the number of extreme precipitation events, which we alleviate by training models specialized towards extreme precipitation events with the imbalanced regression method DenseLoss. While we consider climate MOS, we argue that aspects of ConvMOS may also be beneficial in other domains with geospatial data, such as air pollution modeling or weather forecasts.