# Label-similarity Curriculum Learning

arXiv: https://arxiv.org/abs/1911.06902 · HTML: https://mlsanity.com/paper/1911.06902/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 15 Nov 2019.

## Built on by (most cited first, 11 of 11)

- [Curriculum Learning: A Survey](https://mlsanity.com/paper/2101.10382/) (2021)
- [EfficientTrain: Exploring Generalized Curriculum Learning for Training Visual Backbones](https://mlsanity.com/paper/2211.09703/) (2022)
- [Hybrid Curriculum Learning for Emotion Recognition in Conversation](https://mlsanity.com/paper/2112.11718/) (2021)
- [EfficientTrain++: Generalized Curriculum Learning for Efficient Visual Backbone Training](https://mlsanity.com/paper/2405.08768/) (2024)
- [Learning Rate Curriculum](https://mlsanity.com/paper/2205.09180/) (2022)
- [HiCuLR: Hierarchical Curriculum Learning for Rhetorical Role Labeling of Legal Documents](https://mlsanity.com/paper/2409.18647/) (2024)
- [On adversarial robustness and the use of Wasserstein ascent-descent dynamics to enforce it](https://mlsanity.com/paper/2301.03662/) (2023)
- [CBM: Curriculum by Masking](https://mlsanity.com/paper/2407.05193/) (2024)
- [Learning to Teach with Student Feedback](https://mlsanity.com/paper/2109.04641/) (2021)
- [Fast FixMatch: Faster Semi-Supervised Learning with Curriculum Batch Size](https://mlsanity.com/paper/2309.03469/) (2023)
- [ERNetCL: A novel emotion recognition network in textual conversation based on curriculum learning strategy](https://mlsanity.com/paper/2308.06450/) (2023)

## Abstract (arXiv)

Curriculum learning can improve neural network training by guiding the optimization to desirable optima. We propose a novel curriculum learning approach for image classification that adapts the loss function by changing the label representation. The idea is to use a probability distribution over classes as target label, where the class probabilities reflect the similarity to the true class. Gradually, this label representation is shifted towards the standard one-hot-encoding. That is, in the beginning minor mistakes are corrected less than large mistakes, resembling a teaching process in which broad concepts are explained first before subtle differences are taught. The class similarity can be based on prior knowledge. For the special case of the labels being natural words, we propose a generic way to automatically compute the similarities. The natural words are embedded into Euclidean space using a standard word embedding. The probability of each class is then a function of the cosine similarity between the vector representations of the class and the true label. The proposed label-similarity curriculum learning (LCL) approach was empirically evaluated using several popular deep learning architectures for image classification tasks applied to five datasets including ImageNet, CIFAR100, and AWA2. In all scenarios, LCL was able to improve the classification accuracy on the test data compared to standard training.