{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/sst-self-training-with-self-adaptive-1","title":"SST: Self-training with Self-adaptive Thresholding for Semi-supervised Learning","arxiv_id":"2506.00467","date":"2025-05-31","proceeding":"Information Processing & Management 2025 4","authors":["Shuai Zhao","Heyan Huang","Xinge Li","Xiaokang Chen","Rui Wang"],"abstract":"Neural networks have demonstrated exceptional performance in supervised learning, benefiting from abundant high-quality annotated data. However, obtaining such data in real-world scenarios is costly and labor-intensive. Semi-supervised learning (SSL) offers a solution to this problem. Recent studies, such as Semi-ViT and Noisy Student, which employ consistency regularization or pseudo-labeling, have demonstrated significant achievements. However, they still face challenges, particularly in accurately selecting sufficient high-quality pseudo-labels due to their reliance on fixed thresholds. Recent methods such as FlexMatch and FreeMatch have introduced flexible or self-adaptive thresholding techniques, greatly advancing SSL research. Nonetheless, their process of updating thresholds at each iteration is deemed time-consuming, computationally intensive, and potentially unnecessary. To address these issues, we propose Self-training with Self-adaptive Thresholding (SST), a novel, effective, and efficient SSL framework. SST introduces an innovative Self-Adaptive Thresholding (SAT) mechanism that adaptively adjusts class-specific thresholds based on the model's learning progress. SAT ensures the selection of high-quality pseudo-labeled data, mitigating the risks of inaccurate pseudo-labels and confirmation bias. Extensive experiments demonstrate that SST achieves state-of-the-art performance with remarkable efficiency, generalization, and scalability across various architectures and datasets. Semi-SST-ViT-Huge achieves the best results on competitive ImageNet-1K SSL benchmarks, with 80.7% / 84.9% Top-1 accuracy using only 1% / 10% labeled data. Compared to the fully-supervised DeiT-III-ViT-Huge, which achieves 84.8% Top-1 accuracy using 100% labeled data, our method demonstrates superior performance using only 10% labeled data.","url_abs":"https://arxiv.org/abs/2506.00467v1","url_pdf":"https://arxiv.org/pdf/2506.00467v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[],"methods":[{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"noisy-student","method_name":"Noisy Student"},{"method_slug":"randaugment","method_name":"RandAugment"},{"method_slug":"stochastic-depth","method_name":"Stochastic Depth"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-clothing1m","task":"Image Classification","dataset":"Clothing1M","model":"Super-SST (ViT-Small, 5% Labels)","rank_in_archive_order":2,"of":51,"metrics":{"Accuracy":"75.7%"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-food-101-1","task":"Image Classification","dataset":"Food-101","model":"Semi-SST (ViT-Base, 10% Labels)","rank_in_archive_order":2,"of":11,"metrics":{"Accuracy (%)":"91.5"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-food-101-1","task":"Image Classification","dataset":"Food-101","model":"Super-SST (ViT-Base, 10% Labels)","rank_in_archive_order":3,"of":11,"metrics":{"Accuracy (%)":"91.1"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-food-101-1","task":"Image Classification","dataset":"Food-101","model":"Semi-SST (ViT-Base, 1% Labels)","rank_in_archive_order":6,"of":11,"metrics":{"Accuracy (%)":"86.5"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-food-101-1","task":"Image Classification","dataset":"Food-101","model":"Super-SST (ViT-Base, 1% Labels)","rank_in_archive_order":8,"of":11,"metrics":{"Accuracy (%)":"83.4"},"uses_additional_data":true},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-6","task":"Semi-Supervised Image Classification","dataset":"CIFAR-10, 250 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":1,"of":27,"metrics":{"Percentage error":"2.42±0.13"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-6","task":"Semi-Supervised Image Classification","dataset":"CIFAR-10, 250 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":2,"of":27,"metrics":{"Percentage error":"3.37±0.22"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-7","task":"Semi-Supervised Image Classification","dataset":"CIFAR-10, 40 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":12,"of":21,"metrics":{"Percentage error":"6.35±0.28"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-7","task":"Semi-Supervised Image Classification","dataset":"CIFAR-10, 40 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":17,"of":21,"metrics":{"Percentage error":"9.59"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar","task":"Semi-Supervised Image Classification","dataset":"CIFAR-10, 4000 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":1,"of":49,"metrics":{"Percentage error":"1.41±0.10"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar","task":"Semi-Supervised Image Classification","dataset":"CIFAR-10, 4000 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":2,"of":49,"metrics":{"Percentage error":"1.61±0.18"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-9","task":"Semi-Supervised Image Classification","dataset":"CIFAR-100, 2500 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":1,"of":16,"metrics":{"Percentage error":"16.62±0.28"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-9","task":"Semi-Supervised Image Classification","dataset":"CIFAR-100, 2500 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":2,"of":16,"metrics":{"Percentage error":"18.51±0.36"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-8","task":"Semi-Supervised Image Classification","dataset":"CIFAR-100, 400 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":4,"of":21,"metrics":{"Percentage error":"31.39±0.47"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-8","task":"Semi-Supervised Image Classification","dataset":"CIFAR-100, 400 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":6,"of":21,"metrics":{"Percentage error":"35.50±0.58"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-1","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 1% labeled data","model":"Semi-SST (ViT-Huge)","rank_in_archive_order":5,"of":65,"metrics":{"Top 1 Accuracy":"80.7%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-1","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 1% labeled data","model":"Super-SST (ViT-Huge)","rank_in_archive_order":6,"of":65,"metrics":{"Top 1 Accuracy":"80.3%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-1","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 1% labeled data","model":"Super-SST (ViT-Small distilled)","rank_in_archive_order":9,"of":65,"metrics":{"Top 1 Accuracy":"76.9%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-1","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 1% labeled data","model":"Semi-SST (ViT-Small)","rank_in_archive_order":16,"of":65,"metrics":{"Top 1 Accuracy":"71.4%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-1","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 1% labeled data","model":"Super-SST (ViT-Small)","rank_in_archive_order":20,"of":65,"metrics":{"Top 1 Accuracy":"70.4%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-2","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 10% labeled data","model":"Semi-SST (ViT-Huge)","rank_in_archive_order":4,"of":75,"metrics":{"Top 1 Accuracy":"84.9%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-2","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 10% labeled data","model":"Super-SST (ViT-Huge)","rank_in_archive_order":5,"of":75,"metrics":{"Top 1 Accuracy":"84.8%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-2","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 10% labeled data","model":"Super-SST (ViT-Small distilled)","rank_in_archive_order":10,"of":75,"metrics":{"Top 1 Accuracy":"80.3%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-2","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 10% labeled data","model":"Semi-SST (ViT-Small)","rank_in_archive_order":16,"of":75,"metrics":{"Top 1 Accuracy":"78.6%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-2","task":"Semi-Supervised Image Classification","dataset":"ImageNet - 10% labeled data","model":"Super-SST (ViT-Small)","rank_in_archive_order":17,"of":75,"metrics":{"Top 1 Accuracy":"78.3%"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-stl-1","task":"Semi-Supervised Image Classification","dataset":"STL-10, 1000 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":2,"of":13,"metrics":{"Accuracy":"98.64±0.08"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-stl-1","task":"Semi-Supervised Image Classification","dataset":"STL-10, 1000 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":3,"of":13,"metrics":{"Accuracy":"98.55±0.10"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-2","task":"Semi-Supervised Image Classification","dataset":"cifar-100, 10000 Labels","model":"Semi-SST (ViT-Small)","rank_in_archive_order":1,"of":29,"metrics":{"Percentage error":"13.50±0.14"},"uses_additional_data":false},{"leaderboard":"/sota/semi-supervised-image-classification-on-cifar-2","task":"Semi-Supervised Image Classification","dataset":"cifar-100, 10000 Labels","model":"Super-SST (ViT-Small)","rank_in_archive_order":2,"of":29,"metrics":{"Percentage error":"14.20±0.17"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}