AI Paper Atlas

Research map / Image generation / GANs

GATSBI: Generative Adversarial Training for Simulation-Based Inference

Poornima Ramesh, Jan-Matthis Lueckmann, Jan Boelts, Álvaro Tejero-Cantero, David S. Greenberg, Pedro J. Goncalves, Jakob H. Macke

ICLR 2022 · Image generation › GANs · approach: Generative adversarial network theory · 5 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use GATSBI rather than NLEShallow water model · Shallow water modelcorrelation of posterior mean with ground-truth parameters: 0.88 \pm 0.10Improved
Tune the GAN architecture and training hyperparameters specifically for the Two-Moons model (ablation)GATSBI · Two-MoonsC2ST scoreImproved
Increase the simulation budgetGATSBI · SLCPC2ST scoreImproved
Increase the simulation budgetGATSBI · Two-MoonsC2ST scoreNo clear effect
Use sequential inference instead of amortised GATSBIGATSBI · SBI benchmark problemsC2ST scoreImproved
Use sequential GATSBI instead of amortised GATSBIGATSBI · SBI benchmark problemsC2ST score relative to flow-based methodsHurt
Use GATSBI instead of NPEShallow water model · Shallow water modelposterior calibrationNo clear effect
Use GATSBI instead of NPENoisy camera model · Noisy camera modelcoherence of posterior samplesImproved
Use GATSBI instead of NLEShallow water model · Shallow water modelrecovery of depth-profile shapeImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy