kdb icon indicating copy to clipboard operation
kdb copied to clipboard

AI can help Visualizing the Impacts of Climate Change. This is an open forum to share our work

Results 32 kdb issues
Sort by recently updated
recently updated
newest added

- [ ] Hyperparameters optimization for masker and painter respectively - [ ] Try adding WD data (keep Unity as well) - [ ] Learning rate decaying - [ ]...

priority:high
domain:ml
To Do

Week of July 27th : VICC hackathon for generating images of different extreme weather events Alexia: wildfire (+ burning birds) Tian: extreme heat Méli: better smog Sun : drought (inspo:...

We are currently focusing of flood events, and have made a very simple proof of concept for wildfire events with a CycleGAN. What other weather events could we explore ?...

discussion
priority:low
domain:ml
domain:climate

We need to save segmentation maps of the real dataset (not only categories ground/background, but with the full set of labels vegetation, cars, etc.). A version with 19 Cityscapes labels...

The simulation data here is 19K (11K urban+8.8K suburban). I guess the performance of these two are similar. Please find the links of the experiments below: pretrained: https://www.comet.ml/tianyu-z/advent-ccai/1e62883c43814d40b9da12e13970bacf non-pretrained: https://www.comet.ml/tianyu-z/advent-ccai/b9b25cbeb47843dda48f4ce593797b66...

I transformed the segmentation map of Cityscapes Validation Datasets (frankfurt, lindau, and munster) into Mask images (by changing the value of the road, sidewalks... to 255 and the rest to...

Where would be a good place to keep "general" information on github? For example, I'm thinking about : 1) where the different datasets are on the cluster, what they correspond...

discussion

A list of short-term ideas to try for improving the masker: - [ ] Implement shared DA discriminator operating on mask + latent vector -- Sun - [ ] Use...

priority:high
domain:ml

https://arxiv.org/pdf/1903.02271.pdf

domain:ml
paper