AI Paper Atlas

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

USB: A Unified Semi-supervised Learning Benchmark for Classification

Yidong Wang, Hao Chen, Yue Fan, Wang SUN, Ran Tao, Wenxin Hou, Renjie Wang, Linyi Yang, Zhi Zhou, Lan-Zhe Guo, Heli Qi, Zhen Wu and 10 more

NeurIPS 2022 · Label-efficient and robust learning › Semi-supervised learning · approach: Safe and realistic semi-supervised learning · 43 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
Use pre-trained ViT instead of training ResNet from scratchCV SSL methods · USB CV taskstraining iterations: 1,000k to 200kImproved
Use pre-trained ViT instead of training from scratchCV SSL methods · CV tasksgeneralization performanceImproved
Use pre-trained ViT rather than ViT trained without pre-training (ablation)CV SSL methods · CIFAR-100, 400 labelstest performanceImproved
Use pre-trained ViT rather than WRN-28-8 trained from scratch (ablation)CV SSL methods · CIFAR-100, 400 labelsconvergenceImproved
Use a diverse set of tasks across domains when evaluating SSLSSL algorithm evaluation · CV, NLP, and Audio taskscross-domain ranking consistencyMixed
Use unlabeled data through SSL rather than labeled-only supervised trainingSSL algorithms · USB taskstest performanceMixed
Apply SSL with extremely scarce labeled dataSSL algorithms · USB taskstest performanceHurt
Use ReMixMatch rather than other evaluated SSL methodsCV SSL methods · USB CV tasksmean error rate (%): 26.43Improved
Use AdaMatch rather than other evaluated SSL methodsAudio SSL methods · USB Audio tasksmean error rate (%): 23.61Improved
Use SimMatch rather than other evaluated SSL methodsNLP SSL methods · USB NLP tasksmean error rate (%): 28.34Improved
Use raw waveform rather than Mel spectrogram inputAudio SSL models · Audio tasksgeneralization performanceHurt
Use HuBERT rather than Wav2Vec 2.0Audio SSL methods · Audio tasks other than human-voice Keyword SpottingrobustnessImproved

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