论文概要
ELP 把 linear probing 放进训练过程：周期性重置的分类器在 detached feature 上学习并衡量当前可分性，ELP-SR 再利用 probe 与主分类器之间的差异自适应调整样本正则。
研究问题
主分类器可能已经很自信，但其特征并不一定具有广泛可分性和泛化能力；传统 linear probe 又通常只在训练完成后作为测量工具。
论文贡献
- 提出在线 episodic linear probe，在 detached feature 上训练并周期性重置。
- 把 probe 作为训练期的特征可分性诊断信号。
- 提出 ELP-SR，根据 probe 与主分类器概率分布差异自适应调节样本损失。
证据与评测范围
论文在细粒度、长尾和通用物体识别任务上报告改进；准确数据集、架构与数值应引用 CVPR 实验。
适用范围与局限
ELP 衡量的是线性分类适用性，这很有信息量，但不等于所有表示质量定义；其表现还取决于 probe schedule、分类器设计和任务标签。
Related work 定位
ELP 将表征诊断与正则化连接起来：与离线 linear probe 不同，它把 episodic probe 产生的信号用于同一次训练。
Related Work 表述
Liang 等提出 Episodic Linear Probing，在训练中反复拟合 detached-feature 线性分类器，并利用其与主分类器的差异正则化表示泛化。
这是一段用于说明论文定位的简洁中性表述。
论文官方英文摘要
Understanding network generalization and feature discrimination is an open research problem in visual recognition. Many studies have been conducted to assess the quality of feature representations. One of the simple strategies is to utilize a linear probing classifier to quantitatively evaluate the class accuracy under the obtained features. The typical linear probe is only applied as a proxy at the inference time, but its efficacy in measuring features' suitability for linear classification is largely neglected in training. In this paper, we propose an episodic linear probing (ELP) classifier to reflect the generalization of visual representations in an online manner. ELP is trained with detached features from the network and re-initialized episodically. It demonstrates the discriminability of the visual representations in training. Then, an ELP-suitable Regularization term (ELP-SR) is introduced to reflect the distances of probability distributions between ELP classifier and the main classifier. ELP-SR leverages a re-scaling factor to regularize each sample in training, which modulates the loss function adaptively and encourages the features to be discriminative and generalized. We observe significant improvements in three real-world visual recognition tasks, including fine-grained visual classification, long-tailed visual recognition, and generic object recognition. The performance gains show the effectiveness of our method in improving network generalization and feature discrimination.
摘要仅用于学术识别，版权仍归论文作者或出版方所有，不属于本页 CC BY 许可范围。
依据与出处
| 核对内容 | 论文中的位置 |
|---|---|
| 问题陈述 | Abstract |
| 方法与贡献 | Abstract; ELP and ELP-SR sections |
| 评测结论 | Abstract; fine-grained, long-tailed, and generic recognition experiments |
主要核验来源: IEEE version of record (CVPR 2022 version of record; CVF open-access copy has alternate pagination 9559-9569).
如何引用
科研结论应引用论文本身；只有在复用本站原创解读时才引用本页。
Yuanzhi Liang, Linchao Zhu, Xiaohan Wang, and Yi Yang. “A Simple Episodic Linear Probe Improves Visual Recognition in the Wild.” 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022), 9549-9559. https://doi.org/10.1109/CVPR52688.2022.00934.
复用许可
本页原创解读采用 CC BY 4.0：复用时须署名并链接本页。论文标题、摘要、图表和书目信息不在此许可范围内，仍保留原有权利。 Creative Commons Attribution 4.0 International.