AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Constructing Multiple Tasks for Augmentation: Improving Neural Image Classification with K-Means Features

Tao Gui, Lizhi Qing, Qi Zhang, Jiacheng Ye, Hang Yan, Zichu Fei, Xuanjing Huang

AAAI 2020 · Label-efficient and robust learning › Semi-supervised learning · approach: Self-training theory · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add Meta-MTL with k-means auxiliary tasksSTL classifier · Omniglotaccuracy: 65.72 -> 74.80Improved
Add Meta-MTL auxiliary-task trainingSTL classifier · MNIST, 1% training dataaccuracy: 91.83 -> 94.22Improved
Use the remaining unlabeled data to train embeddings (ablation)Meta-MTL, MNIST · MNIST, 1% training dataaccuracy: 94.22 -> 94.42Improved
Add Meta-MTL with k-means auxiliary tasksSTL classifier · CIFAR-10accuracy: 72.48 -> 75.69Improved
Add Meta-MTL with k-means auxiliary tasksSTL classifier with data augmentation · CIFAR-10, CNN with data augmentationaccuracy: 75.91 -> 79.65Improved
Add Meta-MTL with k-means auxiliary tasksSTL classifier · CIFAR-100, 20 Caccuracy: 55.94 -> 60.39Improved
Add Meta-MTL with k-means auxiliary tasksSTL classifier · CIFAR-100, 100 Caccuracy: 44.19 -> 47.94Improved
Add Meta-MTL to k-means task training (ablation)MTL, k-means labels, miniImageNet DeepCluster embeddings · miniImageNetaccuracy: 37.24 -> 41.07Improved
Use DeepCluster rather than ACAI embeddings (ablation)Meta-MTL, random scaling · miniImageNetaccuracy: 40.43 -> 41.07Improved
Increase auxiliary tasks from T = 4 to T = 12 (ablation)Meta-MTL · CIFAR-10accuracy: 75.02 -> 75.69Improved
Use k-means labels rather than random labels (ablation)Meta-MTL · CIFAR-10, varying number of auxiliary tasksaccuracy: more than 4%Improved
Replace Meta-MTL with ordinary MTL on k-means tasks (ablation)CIFAR-10 · CIFAR-10, varying number of auxiliary tasksaccuracy: more than 8%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy