Paper Reading
0. Solid Ideas
-
Yoshua Bengio重新思考ML的投稿
1.Quantum Computing
- TKDE’16_Relevance Feedback Algorithms Inspired By Quantum Detection
2.Active Learning
3. Ensemble Learning
- 进化集成学习算法综述
- FCS’20 A survey on ensemble learning
- ⭐⭐⭐⭐Access’22 A Survey of Ensemble Learning:过于简单的集成学习入门介绍,公式和引用多,大部分只解决分类问题
-
Artif. Intell. Rev.‘23 A Survey on ensemble learning under the era of deep learning:重视统一的范式,举的例子涉及各个领域,较杂。虽然标题有深度学习,但看完之后感觉更像是一个概念,没有说明怎么和深度学习结合,还得是看具体案例。
Application
- TPMAI’04 Asymmetric Bagging and Random Subspace for Support Vector Machines-Based Relevance Feedback in Image Retrieval :简单的SVM+随机采样,用于消除CBIR中的数据量小或者不平衡。
- Access’17 An_Ensemble_Deep_Learning_Method_for_Vehicle_Type_Classification_on_Visual_Traffic_Surveillance_Sensors:简单的ResNet+过采样解决数据集类别不均匀问题,集成采用max voting policy,还是解决分类问题。
4. Large Language Model(LLM)
- ⭐ ⭐ ⭐ ⭐ ⭐ arxiv 2311’ LLM_Survey_Chinese:很全面的综述,重点讲述最新的模型、训练方法、测试。
- arxiv’21 How much can clip benefit vision-and-language Tasks:用CLIP的视觉编码器完成V&L的任务,实验较完整,理论无拓展。
- ⭐⭐⭐⭐⭐CLIP:Learning Transferable Visual Models From Natural Language Supervision:V&L的开山之作
- ⭐⭐⭐⭐ECCV’22 SLIP:Self-supervision Meets Language-Image Pre-training
- ⭐⭐⭐⭐⭐ICML’22 BLIP Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation:应用效果会比SLIP更佳,论文写的也更清楚些。组合模块+数据清洗后增强
5.Video Retrieval
- ⭐⭐⭐⭐Improving Interpretable Embeddings for Ad-hoc Video Search with Generative Captions and Multi-word Concept Bank:在表达上值得学习,想法好,但是实验结论不是很明显。
待阅读合集
1.机器学习
- 人工智能:现代方法(第3版)
This line appears after every note.
Notes mentioning this note
There are no notes linking to this note.