Timeline for Can a GAN be used for data augmentation?
Current License: CC BY-SA 4.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 2, 2022 at 6:03 | comment | added | keramat | I tried TVAE on iris and diabetes datasets without success here:colab.research.google.com/drive/…. | |
| Feb 1, 2022 at 10:29 | comment | added | keramat | I will try that and report the possible gain. | |
| Feb 1, 2022 at 10:28 | comment | added | usεr11852 |
sdv has GAN- as well as VAE-based models. But yeah, you can always code your own VAE and go forward with that.
|
|
| Feb 1, 2022 at 10:23 | comment | added | keramat | You mean I test it instead of asking!? sdv utilizes VAEs, too? | |
| Feb 1, 2022 at 10:21 | comment | added | usεr11852 |
Yes, sdv is quite easy to use.
|
|
| Feb 1, 2022 at 10:01 | comment | added | keramat | Thanks for your helpful reply. Did you find performance gain with syntactical sample augmentation with VAEs? | |
| Feb 1, 2022 at 9:57 | comment | added | usεr11852 | Yes, I have used it. No, I didn't see any performance gains per se, I used it for model validation tasks. Generally speaking, I found VAEs to be "better" than GANs in a variety of situations (EDA, model validation, pre-training, naive similarity metrics) but maybe that was application dependent. | |
| Feb 1, 2022 at 6:49 | comment | added | keramat | Did you utilize sdv for augmentation? If yes, had it any performance gain? | |
| Dec 23, 2021 at 16:39 | history | answered | usεr11852 | CC BY-SA 4.0 |