Pith. sign in

REVIEW 14 cited by

Unsupervised Data Augmentation for Consistency Training

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1904.12848 v6 pith:J2EM7SJN submitted 2019-04-29 cs.LG cs.AIcs.CLcs.CVstat.ML

Unsupervised Data Augmentation for Consistency Training

classification cs.LG cs.AIcs.CLcs.CVstat.ML
keywords dataexampleslabeledlearningmethodwhenaugmentationconsistency
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce. Common among recent approaches is the use of consistency training on a large amount of unlabeled data to constrain model predictions to be invariant to input noise. In this work, we present a new perspective on how to effectively noise unlabeled examples and argue that the quality of noising, specifically those produced by advanced data augmentation methods, plays a crucial role in semi-supervised learning. By substituting simple noising operations with advanced data augmentation methods such as RandAugment and back-translation, our method brings substantial improvements across six language and three vision tasks under the same consistency training framework. On the IMDb text classification dataset, with only 20 labeled examples, our method achieves an error rate of 4.20, outperforming the state-of-the-art model trained on 25,000 labeled examples. On a standard semi-supervised learning benchmark, CIFAR-10, our method outperforms all previous approaches and achieves an error rate of 5.43 with only 250 examples. Our method also combines well with transfer learning, e.g., when finetuning from BERT, and yields improvements in high-data regime, such as ImageNet, whether when there is only 10% labeled data or when a full labeled set with 1.3M extra unlabeled examples is used. Code is available at https://github.com/google-research/uda.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 14 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Emerging Properties in Self-Supervised Vision Transformers

    cs.CV 2021-04 conditional novelty 8.0

    Self-supervised ViTs show emergent semantic segmentation and 78.3% k-NN accuracy on ImageNet; DINO reaches 80.1% linear evaluation with ViT-Base.

  2. Longformer: The Long-Document Transformer

    cs.CL 2020-04 accept novelty 7.0

    Longformer uses local windowed attention plus task-specific global attention to achieve linear scaling and state-of-the-art results on long-document language modeling, QA, and summarization after pretraining.

  3. A Simple Framework for Contrastive Learning of Visual Representations

    cs.LG 2020-02 accept novelty 7.0

    SimCLR learns visual representations by contrasting augmented views of the same image and reaches 76.5% ImageNet top-1 accuracy with a linear classifier, matching a supervised ResNet-50.

  4. Unsupervised Cross-lingual Representation Learning at Scale

    cs.CL 2019-11 conditional novelty 7.0

    XLM-R, pretrained on 100 languages with 2TB of CommonCrawl data, improves average XNLI accuracy by 14.6 points and MLQA F1 by 13 points over mBERT while matching strong monolingual models on GLUE.

  5. XLNet: Generalized Autoregressive Pretraining for Language Understanding

    cs.CL 2019-06 accept novelty 7.0

    XLNet is a generalized autoregressive pretraining method that learns bidirectional contexts via permutation-based factorization and outperforms BERT on 20 NLP tasks.

  6. Revisiting Feature Prediction for Learning Visual Representations from Video

    cs.CV 2024-02 conditional novelty 6.0

    V-JEPA models trained only on feature prediction from 2 million public videos achieve 81.9% on Kinetics-400, 72.2% on Something-Something-v2, and 77.9% on ImageNet-1K using frozen ViT-H/16 backbones.

  7. Vision Transformers Need Registers

    cs.CV 2023-09 unverdicted novelty 6.0

    Adding register tokens to Vision Transformers eliminates high-norm background artifacts and raises state-of-the-art performance on dense visual prediction tasks.

  8. Graph Star Net for Generalized Multi-Task Learning

    cs.SI 2019-06 unverdicted novelty 6.0

    GraphStar is a new GNN that adds star nodes and relay attention to achieve non-local representations for node, graph, and link tasks, claiming 2-5% gains over prior SOTA on benchmarks.

  9. A Single Rewrite Suffices: Empirical Lessons from Production Skill Description Optimization

    cs.CL 2026-06 unverdicted novelty 5.0

    A single LLM rewrite of skill descriptions using false positive and negative cases matches manual optimization performance in production, with most other pipeline components adding little value.

  10. Invariance-inducing regularization using worst-case transformations suffices to boost accuracy and spatial robustness

    cs.LG 2019-06 unverdicted novelty 5.0

    Invariance-inducing regularization using worst-case transformations reduces relative error by 20% on CIFAR10 transformed examples, improves standard accuracy on SVHN, outperforms equivariant networks, and proves no ac...

  11. Efficient data augmentation using graph imputation neural networks

    stat.ML 2019-06 unverdicted novelty 5.0

    Graph imputation neural networks augment semi-supervised datasets up to 10x by reconstructing heavily damaged samples on a similarity graph, improving over fully-supervised baselines on benchmarks.

  12. Voice Biomarkers for Depression and Anxiety

    cs.LG 2026-05 unverdicted novelty 4.0

    Deep learning models extract content-agnostic voice biomarkers for depression and anxiety from a ~65k-utterance proprietary dataset, achieving 71% sensitivity and specificity when combined with lexical features.

  13. Domain-Specific Query Understanding for Automotive Applications: A Modular and Scalable Approach

    cs.IR 2026-01 unverdicted novelty 4.0

    Decomposing automotive query understanding into a lightweight classification stage followed by specialized entity extraction yields better accuracy and lower latency than joint single-step processing.

  14. Optimizing Small Transformer-Based Language Models for Multi-Label Sentiment Analysis in Short Texts

    cs.CL 2025-09 conditional novelty 4.0

    On SemEval 2025 Task 11 short texts, generated data helped some BERT models, continued pretraining was mixed, and classification head changes barely mattered.