Papers › All Labels Are Not Created Equal: Enhancing Semi-supervision via Label Grouping and Co-training

All Labels Are Not Created Equal: Enhancing Semi-supervision via Label Grouping and Co-training

12 Apr 2021CVPR 2021 1arXiv:2104.05248archive 2025-07-28

Islam Nassar, Samitha Herath, Ehsan Abbasnejad, Wray Buntine, Gholamreza Haffari

Pseudo-labeling is a key component in semi-supervised learning (SSL). It relies on iteratively using the model to generate artificial labels for the unlabeled data to train against. A common property among its various methods is that they only rely on the model's prediction to make labeling decisions without considering any prior knowledge about the visual similarity among the classes. In this paper, we demonstrate that this degrades the quality of pseudo-labeling as it poorly represents visually similar classes in the pool of pseudo-labeled data. We propose SemCo, a method which leverages label semantics and co-training to address this problem. We train two classifiers with two different views of the class labels: one classifier uses the one-hot view of the labels and disregards any potential similarity among the classes, while the other uses a distributed view of the labels and groups potentially similar classes together. We then co-train the two classifiers to learn based on their disagreements. We show that our method achieves state-of-the-art performance across various SSL tasks including 5.6% accuracy improvement on Mini-ImageNet dataset with 1000 labeled examples. We also show that our method requires smaller batch size and fewer training iterations to reach its best performance. We make our code available at https://github.com/islam-nassar/semco.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

islam-nassar/semco officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AllSemi-Supervised Image Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Semi-Supervised Image Classification CIFAR-10, 4000 Labels SemCo (μ=7) Percentage error 3.8±0.08 #4 of 49 Archive leaderboard report
Semi-Supervised Image Classification Mini-ImageNet, 1000 Labels SemCo (μ=3) Accuracy 44.65±0.71 #2 of 3 Archive leaderboard report
Semi-Supervised Image Classification Mini-ImageNet, 1000 Labels SemCo (μ=7) Accuracy 40.65±0.23 #3 of 3 Archive leaderboard report
Semi-Supervised Image Classification Mini-ImageNet, 10000 Labels SemCo (μ=3) Accuracy 58.75±0.76 #2 of 3 Archive leaderboard report
Semi-Supervised Image Classification Mini-ImageNet, 10000 Labels SemCo (μ=7) Accuracy 57.22±0.35 #3 of 3 Archive leaderboard report
Semi-Supervised Image Classification Mini-ImageNet, 4000 Labels SemCo (μ=3) Accuracy 53.99±0.93 #3 of 4 Archive leaderboard report
Semi-Supervised Image Classification Mini-ImageNet, 4000 Labels SemCo (μ=7) Accuracy 50.54±2.20 #4 of 4 Archive leaderboard report
Semi-Supervised Image Classification cifar-100, 10000 Labels SemCo (μ=7) Percentage error 24.45±0.12 #21 of 29 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections