AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Adaptive IMLE for Few-shot Pretraining-free Generative Modelling

Mehran Aghabozorgi, Shichong Peng, Ke Li

ICML 2023 · Learning theory and algorithms › Prediction and decision losses · approach: Theory of neural network sensitivity

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
Replace vanilla IMLE uniform objective with Adaptive IMLE per-example adaptive thresholds (ablation)IMLE few-shot image synthesis · Grumpy Cat, Obama, Panda, Cat, Dog, FFHQ subset (6 few-shot datasets)FID (lower better): e.g. Dog 61.9 -> 43.0 ... Obama 37.4 -> 25.0 ... Panda 8.2 -> 7.6 (mixed in magnitude)Improved
Adaptive IMLE versus FakeCLR, FastGAN, MixDL and EDM baselinesAdaptive IMLE · six few-shot datasets, 256x256FID (lower better): Lower FID than best baseline on all six datasets (e.g. Dog 25.0 vs FakeCLR 44.4)Improved
Adaptive IMLE versus baselines on recall (mode coverage)Adaptive IMLE · six few-shot datasetsRecall: e.g. Grumpy Cat 0.72 vs FakeCLR 0.39; Dog 0.61 vs 0.34Improved
Adaptive IMLE versus baselines on LPIPS backtracking score (lower means better coverage)Adaptive IMLE · six few-shot datasetsLPIPS backtracking: e.g. FFHQ subset 0.014 vs MixDL 0.221Improved

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