论文网址：Vision-Language Models for Vision Tasks: A Survey | IEEE Journals & Magazine | IEEE Xplore
论文Github页面：GitHub - jingyi0000/VLM_survey: Collection of AWESOME vision-language models for vision tasks
英文是纯手打的！论文原文的summarizing and paraphrasing。可能会出现难以避免的拼写错误和语法错误，若有发现欢迎评论指正！文章偏向于笔记，谨慎食用
未完待续
1. 心得
（1）依旧放松一下，以及很久没看TPAMI了，感觉一直很认可TPAMI的质量啊，拜读一下
2. 论文逐段精读
2.1. Abstract
①Existing problems: train DNN for each visual task, which is laborious and time costing
②Content: a) background of VLM in visual task, b) doundations of VLM, c) datasets, d) pretraining, transfer learning and knowledge distillation methods of VLM, e) benchmarks, f) challenges
laborious adj.费力的；辛苦的
2.2. Introduction
①New paradigm: Pre-training (on large scale data w/ or w/o label), Fune-tuning (for specific labelled training data), and Prediction, see (a) and (b):
②Vision-Language Model Pre-training and Zero-shot Prediction which do not need fune-tuning:
③VLM publication number on Google Scholar:
frisbee n.（投掷游戏用的）飞盘；飞碟
2.3. Background
2.3.1. Training Paradigms for Visual Recognition
（1）Traditional Machine Learning and Prediction
①Mostly hand-crafted and lightweight but hard to cope with complex or multi tasks
②Poor scalability
（2）Deep Learning From Scratch and Prediction
①Low speed convergence from scratch
②A mount of labels needed
（3）Supervised Pre-Training, Fine-Tuning and Prediction
①Speed up convergence
（4）Unsupervised Pre-Training, Fine-Tuning & Prediction
①Does not require labelled data
②Beter performance due to larger samples learning
（5）VLM Pre-Training and Zero-Shot Prediction
①Discarding fine-tuning
②Future directions: a) large scale informative image-text data, b) high-capacity models, c) new pre-training objectives
2.3.2. Development of VLMs for Visual Recognition
①3 improvements to VLMs:
2.3.3. Relevant Surveys
①Framework of their review:
2.4. VLM Foundations
2.4.1. Network Architectures
①Number of image-text pairs:
②Features extracted from pairs: , where
with superscript
denotes image sample with
denotes text
③Image encoder and text encoder in DNN: /
④Encoding operation: and
（1）Architectures for Learning Image Features
①CNN-based architectures: such as VGG, ResNet and EfficientNet
②Transformer-base architectures: such as ViT
（2）Architectures for Learning Language Features
①The framework of standard Transformer: 6 blocks in encoder (each with a multi-head attention layer and MLP) and 6 blocks in decoder (each with a multi-head attention layer, a masked multi-head layer and MLP)
2.4.2. VLM Pre-Training Objectives
（1）Contrastive Objectives
①Image Contrastive Learning: close with positive keys and faraway from negative keys in embedding space. For images（实际上作者这里表达得很特殊，他们是说"对于这样的batch size"大小，这是比较贴近代码的表达，如果要概念上的表达其实就看成总共有这么多样本就好）, this loss always be:
where denotes query embedding,
denotes key embeddings,
denotes positive keys in the
-th sample,
denotes temperature hyper-parameter
②Image-Text Contrastive Learning: pull paired embeddings closed and others away:
where denotes contrasting the query image with the text keys,
denotes contrasting the query text with image keys
③Image-Text-Label Contrastive Learning: supervised:
where ,
denotes the class label of
（相当于多增加了一个样本类循环）
（2）Generative Objectives
①Masked Image Modelling: learns cross-patch correlation by masking a set of patches and reconstructing images. The loss usually is:
where denotes masked patches,
denotes unmasked patches（这"|"什么玩意儿啊条件概率吗但是说不通？在不mask的情况下mask的概率？？？怎么感觉反了呢还是我有问题）
②Masked Language Modelling: mask at a specific ratio:
③Masked Cross-Modal Modelling: randomly masks a subset of image patches and a subset of text tokens then reconstruct by unmasked ones:
④Image-to-Text Generation: through image and text pairs to predict text:
where denotes the number of tokens,
is the embedding of the image paired with
（3）Alignment Objectives
①Image-Text Matching: BCE loss:
where measures the alignment probability between the image and text,
when matches otherwise 0
②Region-Word Matching: model local cross-modal correlation in dense scenes:
where denotes a region-word pair,
when matches otherwise 0
2.4.3. VLM Pre-Training Frameworks
①two-tower, two-leg and one-tower pre-training approaches:
2.4.4. Evaluation Setups and Downstream Tasks
（1）Zero-Shot Prediction
①
（2）Linear Probing
2.5. Datasets
2.5.1. Datasets for Pre-Training VLMs
2.5.2. Datasets for VLM Evaluation
2.6. Vision-Language Model Pre-Training
2.6.1. VLM Pre-Training With Contrastive Objectives
（1）Image Contrastive Learning
（2）Image-Text Contrastive Learning
（3）Image-Text-Label Contrastive Learning
（4）Discussion
2.6.2. VLM Pre-Training With Generative Objectives
（1）Masked Image Modelling
（2）Masked Language Modelling
（3）Masked Cross-Modal Modelling
（4）Image-to-Text Generation
（5）Discussion
2.6.3. VLM Pre-Training With Alignment Objectives
（1）Image-Text Matching
（2）Region-Word Matching
（3）Discussion
2.6.4. Summary and Discussion
2.7. VLM Transfer Learning
2.7.1. Motivation of Transfer Learning
2.7.2. Common Setup of Transfer Learning
2.7.3. Common Transfer Learning Methods
（1）Transfer Via Prompt Tuning
（2）Transfer Via Feature Adaptation
（3）Other Transfer Methods
2.7.4. Summary and Discussion
2.8. VLM Knowledge Distillation
2.8.1. Motivation of Distilling Knowledge From VLMs
2.8.2. Common Knowledge Distillation Methods
（1）Knowledge Distillation for Object Detection
（2）Knowledge Distillation for Semantic Segmentation
2.8.3. Summary and Discussion
2.9. Performance Comparison
2.9.1. Performance of VLM Pre-Training
2.9.2. Performance of VLM Transfer Learning
2.9.3. Performance of VLM Knowledge Distillation
2.9.4. Summary
2.10. Future Directions
2.11. Conclusion
- 知识补充
4. Reference
Zhang, J. et al. (2024) Vision-Language Models for Vision Tasks: A Survey. TPAMI, 46(8): 5625-5644. doi: 10.1109/TPAMI.2024.3369699