no code implementations • 25 Mar 2024 • Busra Asan, Abdullah Akgül, Alper Unal, Melih Kandemir, Gozde Unal
Seasonal forecasting is a crucial task when it comes to detecting the extreme heat and colds that occur due to climate change.
no code implementations • 11 Mar 2024 • Batuhan Cengiz, Mert Gulsen, Yusuf H. Sahin, Gozde Unal
Due to the wide application area of point clouds and the recent advancements in deep neural networks, studies focusing on robust classification of the 3D point cloud data emerged.
1 code implementation • 11 Mar 2024 • Mert Gulsen, Batuhan Cengiz, Yusuf H. Sahin, Gozde Unal
A typical way to assess a model's robustness is through adversarial attacks, where test-time examples are generated based on gradients to deceive the model.
1 code implementation • 9 Oct 2023 • Gulcin Baykal, Melih Kandemir, Gozde Unal
We evidentially monitor the significance of attaining the probability distribution over the codebook embeddings, in contrast to softmax usage.
1 code implementation • 4 Jul 2023 • Gulcin Baykal, Halil Faruk Karagoz, Taha Binhuraib, Gozde Unal
Diffusion models are generative models that have shown significant advantages compared to other generative models in terms of higher generation quality and more stable training.
no code implementations • 2 Apr 2023 • Halil Faruk Karagoz, Gulcin Baykal, Irem Arikan Eksi, Gozde Unal
The fine-tuned diffusion model is trained with this newly created dataset, and its results are compared with the baseline models visually and numerically.
1 code implementation • 7 Mar 2023 • V. Bugra Yesilkaynak, Emine Dari, Alican Mertan, Gozde Unal
We show that our method is able to accurately learn a representation of the incorporated positive rank order, which is not only consistent with the ground truth but also proportional to the underlying information.
no code implementations • 21 Feb 2023 • Alper Unal, Busra Asan, Ismail Sezen, Bugra Yesilkaynak, Yusuf Aydin, Mehmet Ilicak, Gozde Unal
Three different setups (CMIP6; CMIP6 + elevation; CMIP6 + elevation + ERA5 finetuning) were used with both UNet and UNet++ algorithms resulting in six different models.
1 code implementation • 8 Dec 2022 • Emine Dari, V. Bugra Yesilkaynak, Alican Mertan, Gozde Unal
Multi-label ranking maps instances to a ranked set of predicted labels from multiple possible classes.
no code implementations • 23 Oct 2022 • Sevgi Altun, Mustafa Cem Gunes, Yusuf H. Sahin, Alican Mertan, Gozde Unal, Mine Ozkar
This study integrates artificial intelligence and computational design tools to extract information from architectural heritage.
1 code implementation • 22 Jun 2022 • Atahan Ozer, Kadir Burak Buldu, Abdullah Akgül, Gozde Unal
Federated Learning enables multiple data centers to train a central model collaboratively without exposing any confidential data.
no code implementations • 2 Mar 2022 • Abdullah Akgül, Gozde Unal, Melih Kandemir
The learning model is aware when a new mode appears, but it cannot access the true modes of individual training sequences.
no code implementations • 9 Aug 2021 • Gozdenur Demir, Asli Cekmis, Vahit Bugra Yesilkaynak, Gozde Unal
Visual design is associated with the use of some basic design elements and principles.
1 code implementation • 6 Aug 2021 • Ufuk Demir, Atahan Ozer, Yusuf H. Sahin, Gozde Unal
However, there are two drawbacks of the approach: most of the edges in the graph are assigned randomly and the GCN is trained independently from the segmentation network.
2 code implementations • ICLR 2022 • Melih Kandemir, Abdullah Akgül, Manuel Haussmann, Gozde Unal
A probabilistic classifier with reliable predictive uncertainties i) fits successfully to the target domain data, ii) provides calibrated class probabilities in difficult regions of the target domain (e. g.\ class overlap), and iii) accurately identifies queries coming out of the target domain and rejects them.
no code implementations • 13 Apr 2021 • Alican Mertan, Damien Jade Duff, Gozde Unal
We review solutions to the problem of depth estimation, arguably the most important subtask in scene understanding.
1 code implementation • 8 Dec 2020 • Yusuf H. Sahin, Alican Mertan, Gozde Unal
Learning new representations of 3D point clouds is an active research area in 3D vision, as the order-invariant point cloud structure still presents challenges to the design of neural network architectures.
Ranked #18 on 3D Part Segmentation on ShapeNet-Part
1 code implementation • 3 Nov 2020 • Gulcin Baykal, Furkan Ozcelik, Gozde Unal
Lastly, we show the contribution of the self-supervision tasks to the GAN training on the loss landscape and present that the effects of these tasks may not be cooperative to the adversarial training in some settings.
no code implementations • 14 Oct 2020 • Alican Mertan, Damien Jade Duff, Gozde Unal
To this end, we have introduced a listwise ranking loss borrowed from ranking literature, weighted ListMLE, to the relative depth estimation problem.
no code implementations • 14 Oct 2020 • Alican Mertan, Yusuf Huseyin Sahin, Damien Jade Duff, Gozde Unal
We propose a new approach for the problem of relative depth estimation from a single image.
1 code implementation • 14 Sep 2020 • Vahit Bugra Yesilkaynak, Yusuf H. Sahin, Gozde Unal
Deep neural network training without pre-trained weights and few data is shown to need more training iterations.
Ranked #1 on Semantic Segmentation on Cityscapes VIPriors subset
1 code implementation • 30 Jun 2020 • Furkan Ozcelik, Ugur Alganci, Elif Sertel, Gozde Unal
Convolutional Neural Networks (CNN)-based approaches have shown promising results in pansharpening of satellite images in recent years.
1 code implementation • 15 Jun 2020 • Gulcin Baykal, Gozde Unal
Generative Adversarial Networks (GANs) triggered an increased interest in problem of image generation due to their improved output image quality and versatility for expansion towards new methods.
no code implementations • 21 May 2019 • M. Jorge Cardoso, Aasa Feragen, Ben Glocker, Ender Konukoglu, Ipek Oguz, Gozde Unal, Tom Vercauteren
This compendium gathers all the accepted extended abstracts from the Second International Conference on Medical Imaging with Deep Learning (MIDL 2019), held in London, UK, 8-10 July 2019.
no code implementations • 12 Apr 2018 • Sahin Olut, Yusuf Huseyin Sahin, Ugur Demir, Gozde Unal
To that end, we incorporate steerable filter responses of the generated and reference images inside a Huber function loss term.
no code implementations • 20 Mar 2018 • Ugur Demir, Gozde Unal
Area of image inpainting over relatively large missing regions recently advanced substantially through adaptation of dedicated deep neural networks.
no code implementations • 31 Dec 2017 • Ugur Demir, Gozde Unal
Then the second network modifies the repaired image to cancel the noise introduced by the first network.