AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

The Label Imitation Game: Turing Test Network for Zero-Shot Pseudo-Label Pruning

Brent Griffin, Jason Corso

ECCV 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Transductive and semantic-dictionary zero-shot

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
prune pseudo-labels with zero-shot TTNVLM pseudo-labels (AL) · Overall pseudo-labels, 4 datasetsF1: 0.362 -> 0.376Improved
fine-tune TTN on detection data (TTN_D)TTN · Overall pseudo-labels, 4 datasetsF1: 0.376 -> 0.387Improved
TTN pruning on min-F1 (hardest) classesAL · Min-F1 classes meanF1: 0.133 -> 0.170Improved
TTN_D pruning on min-F1 classesAL · Min-F1 classes meanF1: 0.133 -> 0.192Improved
prune pseudo-labels (zero-shot TTN) on LVISAL · LVIS pseudo-labelsF1: 0.078 -> 0.079No clear effect
train detector on TTN-pruned labelsYOLO11n on AL labels · Overall downstreammAP50: 0.350 -> 0.355Improved
train detector on TTN_D-pruned labelsYOLO11n on AL labels · Overall vulnerable classesmAP50: 0.155 -> 0.171Improved
pruning reduces recallAL · Overall pseudo-labelsrecall: 0.453 -> 0.437Hurt
use CLIP kNN relabeling instead of pruningAL · Pseudo-label evaluationF1: 0.362 -> 0.275Hurt
use TTN tokenizer embeddings for kNN instead of CLIP (ablation)kNN prune · Pseudo-label evaluationF1: 0.366 -> 0.373Improved
use ground-truth reference set instead of self-referential (ablation)TTN · Pseudo-label evaluationF2: 0.386 -> 0.395Improved
shrink self-referential reference set to <=100 (ablation)TTN · Pseudo-label evaluationF2: 0.386 -> 0.381Hurt

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