Work

Marginal Calibration Masks Generator Differences in Held-Out Evaluation of Multivariate Time Series, with an Application to ERA5 Weather Sequences

Year: 2026

Type: software

Abstract: A multi-discriminator GAN for synthetic multivariate weather sequences and a held-out re-evaluation that overturns most of its own architecture claims. Code, ablation results, and figures for the StatD2GAN v2 campaign (5 locations, 9 ablation arms x 5 seeds, 2 baselines x 25 runs, held-out evaluation protocol).

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.22996771

Authors: Özaytaç, Mustafa, Karadağ Ataş, Özge

Institution: Hacettepe University

Language: English

Topic: Generative Adversarial Networks and Image Synthesis

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7214553635

Cites:

Cited by:

Related to: 10