Pith. sign in

REVIEW 10 cited by

Unsupervised Prompt Learning for Vision-Language Models

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2204.03649 v2 pith:H5QB6QZP submitted 2022-04-07 cs.CV

Tony Huang , Jack Chu , Fangyun Wei This is my paper
classification cs.CV
keywords promptlearningmodelsdatasetsvision-languageengineeringunsupervisedavoid
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Contrastive vision-language models like CLIP have shown great progress in transfer learning. In the inference stage, the proper text description, also known as prompt, needs to be carefully designed to correctly classify the given images. In order to avoid laborious prompt engineering, recent works such as CoOp, CLIP-Adapter and Tip-Adapter propose to adapt vision-language models for downstream image recognition tasks on a small set of labeled data. Though promising improvements are achieved, requiring labeled data from the target datasets may restrict the scalability. In this paper, we explore a different scenario, in which the labels of the target datasets are unprovided, and we present an unsupervised prompt learning (UPL) approach to avoid prompt engineering while simultaneously improving transfer performance of CLIP-like vision-language models. As far as we know, UPL is the first work to introduce unsupervised learning into prompt learning. Experimentally, our UPL outperforms original CLIP with prompt engineering on ImageNet as well as other 10 datasets. An enhanced version of UPL is even competitive with the 8-shot CoOp and the 8-shot TIP-Adapter on most datasets. Code and models are available at https://github.com/tonyhuang2022/UPL.

Discussion (0). Sign in to comment.

Forward citations

Cited by 10 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Bi-CoG: Bi-Consistency-Guided Self-Training for Vision-Language Models

    cs.LG 2025-10 unverdicted novelty 6.0 of 10

    Bi-CoG improves semi-supervised fine-tuning of VLMs by combining inter-model voting, intra-model consistency, and an error-aware pseudo-label budget.

  2. C-GAP: Class-Aware and Online Prompting Improves Vision-Language Models on Imbalanced Classes

    cs.CV 2026-07 conditional novelty 5.5 of 10

    Iterative LLM caption refinement guided by minority-class AP@0.5 lifts rare-object detection on frozen open-vocabulary detectors without labels or weight updates.

  3. USE: A Unified Self-Ensembling Framework for Test-Time Prompt Tuning

    cs.CV 2026-07 accept novelty 5.5 of 10

    Adaptive self-ensembling of weak and strong views yields more reliable pseudo-labels for test-time prompt tuning of CLIP and unifies training with inference.

  4. $\Delta \mathrm{Energy}$: Optimizing Energy Change During Vision-Language Alignment Improves both OOD Detection and OOD Generalization

    cs.CV 2025-10 reject novelty 5.0 of 10

    ΔEnergy, an energy-change OOD score for CLIP, and its EBM fine-tuning loss simultaneously improve OOD detection and covariate-shift generalization.

  5. Adapting Vision-Language Models Without Labels: A Comprehensive Survey

    cs.LG 2025-08 conditional novelty 5.0 of 10

    A survey that organizes unsupervised vision-language model adaptation by unlabeled-data availability into four paradigms: data-free transfer, domain transfer, episodic test-time, and online test-time adaptation.

  6. Prototype-Guided Pseudo-Labeling with Neighborhood-Aware Consistency for Unsupervised Adaptation

    cs.LG 2025-07 conditional novelty 5.0 of 10

    A prototype- and neighbor-aware pseudo-labeling framework that reports improved unsupervised CLIP adaptation accuracy on 11 image classification benchmarks.

  7. Towards Fine-Grained Adaptation of CLIP via a Self-Trained Alignment Score

    cs.CV 2025-07 conditional novelty 5.0 of 10

    FAIR adapts CLIP to unlabeled fine-grained domains by pseudo-labeling through a learned alignment score between localized image crops and learnable class anchors, claiming an average 2.78% top-1 gain over SOTA on 13 datasets.

  8. Progressive Homeostatic and Plastic Prompt Tuning for Audio-Visual Multi-Task Incremental Learning

    cs.AI 2025-07 reject novelty 4.0 of 10

    A three-stage prompt-tuning method for audio-visual multi-task incremental learning is proposed, reporting state-of-the-art results on AVE, AVVP, AVS, and AVQA, with caveats about its evaluation metric and ablations.

  9. Advancing Reliable Test-Time Adaptation of Vision-Language Models under Visual Variations

    cs.CV 2025-07 conditional novelty 4.0 of 10

    ReTA improves CLIP's test-time adaptation by reweighting sample selection with prediction consistency and adapting class decision boundaries via Gaussian text embeddings.

  10. Vision-Language Models display a strong gender bias

    cs.CV 2025-08 reject novelty 3.0 of 10

    Using cosine similarity in CLIP embedding space, the paper finds that male and female face sets are differentially associated with occupation and activity statements across all four tested models.

Pith tools