AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

PSP: Prompt-Guided Self-Training Sampling Policy for Active Prompt Learning

Sen Tao, Kaiduo Feng, Jiawei Liu, Peng Zeng, Yongchao Xu, Yufei Zheng, Zheng-Jun Zha

ICLR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Prompt learning for CLIP

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 PCB sampling with PSP (VSSP + UST)PCB (AS) with BADGE · Average over 7 datasetsfinal accuracy: 74.36 -> 76.87Improved
PSP vs PCB (AS) on large-scale categoriesBADGE + PCB (AS) · Stanford Carsfinal accuracy: 70.70 -> 73.84Improved
PSP vs best baseline on Flowers102 and Caltech101BADGE + PCB (AS) · Flowers102final accuracy: 96.71 -> 96.35No clear effect
add vectorized soft actor-critic sampling policy (VSSP) (ablation)PCB + UST · DTD/Pets/EuroSAT/Aircraft avgfinal accuracy: 67.26 -> 68.52Improved
add uncertainty-augmented self-training (UST) pseudo-labels (ablation)PCB + VSSP · DTD/Pets/EuroSAT/Aircraft avgfinal accuracy: 66.42 -> 68.52Improved
use state with classification-score gradient embeddings instead of features only (ablation)VSSP · DTDfinal accuracy: 63.36 -> 65.66Improved
set beta=0.7 vs beta=0.0 in hybrid reward (ablation)PSP · DTDfinal accuracy: 63.89 -> 65.66Improved
replace UPL* pseudo-labeling with UST (ablation)prompt learning on CLIP · DTDfinal accuracy: 55.08 -> 62.65Improved
replace XPL with UST (ablation)prompt learning on CLIP · DTDfinal accuracy: 62.29 -> 62.65No clear effect
swap CLIP for SigLIP, PSP vs PCBactive prompt learning · avg of 4 datasetsfinal accuracy: 48.82 -> 59.39Improved
PSP vs CoOp/CoCoOp with same real labelsprompt learning · Average over 7 datasetsfinal accuracy: 69.11 -> 76.87 (CoOp)Improved

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