Unsupervised Data Augmentation for Consistency Training

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.

NeurIPS 2020

References in corpus (36)

• Distributed Representations of Words and Phrases and their Compositionality
• Distilling the Knowledge in a Neural Network
• Semi-Supervised Classification with Graph Convolutional Networks
• mixup: Beyond Empirical Risk Minimization
• SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
• Improved Regularization of Convolutional Neural Networks with Cutout
• Wide Residual Networks
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
• Semi-Supervised Learning with Deep Generative Models
• Temporal Ensembling for Semi-Supervised Learning
• Interpolation Consistency Training for Semi-Supervised Learning
• Data-Efficient Image Recognition with Contrastive Predictive Coding
• AutoAugment: Learning Augmentation Policies from Data
• MixMatch: A Holistic Approach to Semi-Supervised Learning
• Semi-Supervised Learning with Ladder Networks
• Semi-supervised Sequence Learning
• QANet: Combining Local Convolution with Global Self-Attention for Reading Comprehension
• Learning with Pseudo-Ensembles
• Shake-Shake regularization
• Learning Discrete Representations via Information Maximizing Self-Augmented Training
• Revisiting Semi-Supervised Learning with Graph Embeddings
• Auxiliary Deep Generative Models
• Adversarial Training Methods for Semi-Supervised Text Classification
• ShakeDrop Regularization for Deep Residual Learning
• Are Labels Required for Improving Adversarial Robustness?
• There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average
• Understanding Back-Translation at Scale
• Data augmentation instead of explicit regularization
• Adversarially Robust Generalization Just Requires More Unlabeled Data
• Selfie: Self-supervised Pretraining for Image Embedding
• Unlabeled Data Improves Adversarial Robustness
• Semi-Supervised QA with Generative Domain-Adaptive Nets
• Semi-Supervised Sequence Modeling with Cross-View Training
• Semi-Supervised Learning by Label Gradient Alignment
• Sequence to Sequence Mixture Model for Diverse Machine Translation
• SwitchOut: an Efficient Data Augmentation Algorithm for Neural Machine Translation

Cited by in corpus (295)

• A Simple Framework for Contrastive Learning of Visual Representations
• Language Models are Few-Shot Learners
• FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
• Longformer: The Long-Document Transformer
• Unsupervised Learning of Visual Features by Contrasting Cluster Assignments
• XLNet: Generalized Autoregressive Pretraining for Language Understanding
• Recent advances and clinical applications of deep learning in medical image analysis
• Data-Efficient Image Recognition with Contrastive Predictive Coding
• Graph Contrastive Learning with Augmentations
• A Comprehensive Survey of Image Augmentation Techniques for Deep Learning
• AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
• Big Self-Supervised Models are Strong Semi-Supervised Learners
• A Survey on Data Augmentation for Text Classification
• Rethinking Pre-training and Self-training
• Deep Learning Based Text Classification: A Comprehensive Review
• Data Augmentation Approaches in Natural Language Processing: A Survey
• Unleashing the potential of prompt engineering for large language models
• A Simple Semi-Supervised Learning Framework for Object Detection
• Large-Scale Adversarial Training for Vision-and-Language Representation Learning
• An Overview of Deep Semi-Supervised Learning
• A Comprehensive Survey on Test-Time Adaptation under Distribution Shifts
• Reinforcement Learning with Augmented Data
• Self-training with Noisy Student improves ImageNet classification
• Improved Noisy Student Training for Automatic Speech Recognition
• A survey on Semi-, Self- and Unsupervised Learning for Image Classification
• PseudoSeg: Designing Pseudo Labels for Semantic Segmentation
• Pushing the Limits of Semi-Supervised Learning for Automatic Speech Recognition
• Are Labels Required for Improving Adversarial Robustness?
• Ten Years of Generative Adversarial Nets (GANs): A survey of the state-of-the-art
• ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
• Data Augmentation using Pre-trained Transformer Models
• Revisiting Self-Training for Neural Sequence Generation
• One Million Scenes for Autonomous Driving: ONCE Dataset
• Image Augmentations for GAN Training
• Generative AI for Synthetic Data Across Multiple Medical Modalities: A Systematic Review of Recent Developments and Challenges
• Memory-aware curriculum federated learning for breast cancer classification
• Self-Training: A Survey
• Generative Data Augmentation for Commonsense Reasoning
• A Simple but Tough-to-Beat Data Augmentation Approach for Natural Language Understanding and Generation
• Fake News in Sheep's Clothing: Robust Fake News Detection Against LLM-Empowered Style Attacks
• Contrast to Divide: Self-Supervised Pre-Training for Learning with Noisy Labels
• Distribution Aligning Refinery of Pseudo-label for Imbalanced Semi-supervised Learning
• Theoretical Analysis of Self-Training with Deep Networks on Unlabeled Data
• Selfie: Self-supervised Pretraining for Image Embedding
• Towards Label-efficient Automatic Diagnosis and Analysis: A Comprehensive Survey of Advanced Deep Learning-based Weakly-supervised, Semi-supervised and Self-supervised Techniques in Histopathological Image Analysis
• Few-Shot Text Generation with Pattern-Exploiting Training
• Self-Supervised Vision Transformers for Malware Detection
• Robust Pre-Training by Adversarial Contrastive Learning
• Unlabeled Data Improves Adversarial Robustness
• Ontology-enhanced Prompt-tuning for Few-shot Learning
• Augmenting Low-Resource Text Classification with Graph-Grounded Pre-training and Prompting
• A Systematic Survey of Regularization and Normalization in GANs
• Supervised Contrastive Learning for Pre-trained Language Model Fine-tuning
• Synthetic Data Generation Using Large Language Models: Advances in Text and Code
• Towards Utilizing Unlabeled Data in Federated Learning: A Survey and Prospective
• Affinity and Diversity: Quantifying Mechanisms of Data Augmentation
• Data-Efficient Reinforcement Learning with Self-Predictive Representations
• Understanding and Mitigating the Tradeoff Between Robustness and Accuracy
• Consistency Regularization for Generative Adversarial Networks
• Structured Consistency Loss for semi-supervised semantic segmentation
• Network Randomization: A Simple Technique for Generalization in Deep Reinforcement Learning
• Towards Zero-Label Language Learning
• Semi-Supervised Neural Architecture Search
• Self-training Improves Pre-training for Natural Language Understanding
• Model Patching: Closing the Subgroup Performance Gap with Data Augmentation
• Using Prior Knowledge to Guide BERT's Attention in Semantic Textual Matching Tasks
• Provable Guarantees for Self-Supervised Deep Learning with Spectral Contrastive Loss
• Uncertainty-aware Self-training for Text Classification with Few Labels
• SemiFL: Semi-Supervised Federated Learning for Unlabeled Clients with Alternate Training
• Rethinking Distributional Matching Based Domain Adaptation
• CReST: A Class-Rebalancing Self-Training Framework for Imbalanced Semi-Supervised Learning
• Adaptive Self-training for Few-shot Neural Sequence Labeling
• Graph Star Net for Generalized Multi-Task Learning
• Feature-aware conditional GAN for category text generation
• Unsupervised Domain Adaptation for RF-based Gesture Recognition
• Regularizing Class-wise Predictions via Self-knowledge Distillation
• MixKD: Towards Efficient Distillation of Large-scale Language Models
• Towards a Robust and Trustworthy Machine Learning System Development: An Engineering Perspective
• CoDA: Contrast-enhanced and Diversity-promoting Data Augmentation for Natural Language Understanding
• Machine Learning with Confidential Computing: A Systematization of Knowledge
• Parting with Illusions about Deep Active Learning
• SelfMatch: Combining Contrastive Self-Supervision and Consistency for Semi-Supervised Learning
• Improved Consistency Regularization for GANs
• Learning with Instance-Dependent Label Noise: A Sample Sieve Approach
• A Survey of Deep Learning Approaches for OCR and Document Understanding
• Semi-Supervised Building Footprint Generation with Feature and Output Consistency Training
• Unsupervised Cross-lingual Adaptation for Sequence Tagging and Beyond
• Article Reranking by Memory-Enhanced Key Sentence Matching for Detecting Previously Fact-Checked Claims
• MixText: Linguistically-Informed Interpolation of Hidden Space for Semi-Supervised Text Classification
• Teacher-Student chain for efficient semi-supervised histology image classification
• SemiFed: Semi-supervised Federated Learning with Consistency and Pseudo-Labeling
• Graph Random Neural Network for Semi-Supervised Learning on Graphs
• Regularizing Deep Networks with Semantic Data Augmentation
• OpenMatch: Open-set Consistency Regularization for Semi-supervised Learning with Outliers
• SemiCurv: Semi-Supervised Curvilinear Structure Segmentation
• Visual Imitation Made Easy
• AdaMatch: A Unified Approach to Semi-Supervised Learning and Domain Adaptation
• GraphCrop: Subgraph Cropping for Graph Classification
• The NTT DCASE2020 Challenge Task 6 system: Automated Audio Captioning with Keywords and Sentence Length Estimation
• Unsupervised Semantic Aggregation and Deformable Template Matching for Semi-Supervised Learning
• Semi-supervised Left Atrium Segmentation with Mutual Consistency Training
• Adaptive Betweenness Clustering for Semi-Supervised Domain Adaptation
• LM-CPPF: Paraphrasing-Guided Data Augmentation for Contrastive Prompt-Based Few-Shot Fine-Tuning
• Not All Unlabeled Data are Equal: Learning to Weight Data in Semi-supervised Learning
• Zero-shot Entity Linking with Efficient Long Range Sequence Modeling
• Iterative self-transfer learning: A general methodology for response time-history prediction based on small dataset
• On Data Augmentation for Extreme Multi-label Classification
• Video Unsupervised Domain Adaptation with Deep Learning: A Comprehensive Survey
• Improving Robustness of Task Oriented Dialog Systems
• Improving Semi-supervised Federated Learning by Reducing the Gradient Diversity of Models
• Deep Reinforcement Learning with Robust and Smooth Policy
• On Data-Augmentation and Consistency-Based Semi-Supervised Learning
• Duplex Conversation: Towards Human-like Interaction in Spoken Dialogue Systems
• Self-supervised Pre-training with Hard Examples Improves Visual Representations
• Semi-supervised learning using teacher-student models for vocal melody extraction
• Beyond Cats and Dogs: Semi-supervised Classification of fuzzy labels with overclustering
• Fuzzy Overclustering: Semi-Supervised Classification of Fuzzy Labels with Overclustering and Inverse Cross-Entropy
• A New Semi-supervised Learning Benchmark for Classifying View and Diagnosing Aortic Stenosis from Echocardiograms
• Active Self-Semi-Supervised Learning for Few Labeled Samples
• Bridging the domain gap in cross-lingual document classification
• Neural Analysis and Synthesis: Reconstructing Speech from Self-Supervised Representations
• Consistency-based Semi-supervised Active Learning: Towards Minimizing Labeling Cost
• KeepAugment: A Simple Information-Preserving Data Augmentation Approach
• FedCon: A Contrastive Framework for Federated Semi-Supervised Learning
• Multi-organ segmentation: a progressive exploration of learning paradigms under scarce annotation
• Monocular 3D Detection with Geometric Constraints Embedding and Semi-supervised Training
• Robust Mutual Learning for Semi-supervised Semantic Segmentation
• On the Generalization Effects of Linear Transformations in Data Augmentation
• Influence-guided Data Augmentation for Neural Tensor Completion
• Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimation
• Proposal Learning for Semi-Supervised Object Detection
• Effects of Word-frequency based Pre- and Post- Processings for Audio Captioning
• Momentum^2 Teacher: Momentum Teacher with Momentum Statistics for Self-Supervised Learning
• Semi-Supervised Models via Data Augmentationfor Classifying Interactive Affective Responses
• Ego-Vehicle Action Recognition based on Semi-Supervised Contrastive Learning
• FedSiam: Towards Adaptive Federated Semi-Supervised Learning
• CO2: Consistent Contrast for Unsupervised Visual Representation Learning
• 3DIoUMatch: Leveraging IoU Prediction for Semi-Supervised 3D Object Detection
• Neural Semi-supervised Learning for Text Classification Under Large-Scale Pretraining
• Improving Dialogue Breakdown Detection with Semi-Supervised Learning
• Improving BERT with Self-Supervised Attention
• XtremeDistilTransformers: Task Transfer for Task-agnostic Distillation
• Combat COVID-19 Infodemic Using Explainable Natural Language Processing Models
• Cycle Self-Training for Semi-Supervised Object Detection with Distribution Consistency Reweighting
• Adaptive Domain Generalization via Online Disagreement Minimization
• Semantic keypoint-based pose estimation from single RGB frames
• Exponential Moving Average Normalization for Self-supervised and Semi-supervised Learning
• SST: Self-training with Self-adaptive Thresholding for Semi-supervised Learning
• Learning to Impute: A General Framework for Semi-supervised Learning
• Invariance-inducing regularization using worst-case transformations suffices to boost accuracy and spatial robustness
• Semi-Supervised Medical Image Segmentation with Co-Distribution Alignment
• Unsupervised Domain Adaptation of Contextualized Embeddings for Sequence Labeling
• Semi-Supervised Learning with Normalizing Flows
• The State of Knowledge Distillation for Classification
• Semi-supervised Viewpoint Estimation with Geometry-aware Conditional Generation
• Robust Visual Domain Randomization for Reinforcement Learning
• Semi-supervised ASR by End-to-end Self-training
• LearnDA: Learnable Knowledge-Guided Data Augmentation for Event Causality Identification
• Self-Supervision, Remote Sensing and Abstraction: Representation Learning Across 3 Million Locations
• OpenCoS: Contrastive Semi-supervised Learning for Handling Open-set Unlabeled Data
• Revisiting consistency for semi-supervised semantic segmentation
• A Simple Baseline for Semi-supervised Semantic Segmentation with Strong Data Augmentation
• Sinkhorn Label Allocation: Semi-Supervised Classification via Annealed Self-Training
• A Three-Stage Self-Training Framework for Semi-Supervised Semantic Segmentation
• AutoConv: Automatically Generating Information-seeking Conversations with Large Language Models
• Self-supervised Document Clustering Based on BERT with Data Augment
• A comparative study of semi- and self-supervised semantic segmentation of biomedical microscopy data
• Sentiment Analysis Using Averaged Weighted Word Vector Features
• An analysis of over-sampling labeled data in semi-supervised learning with FixMatch
• SeqMix: Augmenting Active Sequence Labeling via Sequence Mixup
• Back-Translated Task Adaptive Pretraining: Improving Accuracy and Robustness on Text Classification
• Boundary-Refined Prototype Generation: A General End-to-End Paradigm for Semi-Supervised Semantic Segmentation
• Unsupervised Pre-training for Person Re-identification
• HiddenCut: Simple Data Augmentation for Natural Language Understanding with Better Generalization
• Robust Training of Social Media Image Classification Models for Rapid Disaster Response
• Negative sampling in semi-supervised learning
• FedSEAL: Semi-Supervised Federated Learning with Self-Ensemble Learning and Negative Learning
• Re-distributing Biased Pseudo Labels for Semi-supervised Semantic Segmentation: A Baseline Investigation
• Semi-Supervised Crowd Counting via Self-Training on Surrogate Tasks
• Data-Uncertainty Guided Multi-Phase Learning for Semi-Supervised Object Detection
• An Effective, Robust and Fairness-aware Hate Speech Detection Framework
• Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach
• SHOT-VAE: Semi-supervised Deep Generative Models With Label-aware ELBO Approximations
• Improving Speaker Identification for Shared Devices by Adapting Embeddings to Speaker Subsets
• A Small Claims Court for the NLP: Judging Legal Text Classification Strategies With Small Datasets
• Towards Principled Disentanglement for Domain Generalization
• GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference
• Bridge the Domain Gap Between Ultra-wide-field and Traditional Fundus Images via Adversarial Domain Adaptation
• BiaSwap: Removing dataset bias with bias-tailored swapping augmentation
• Improving Molecular Design by Stochastic Iterative Target Augmentation
• How Useful Are the Machine-Generated Interpretations to General Users? A Human Evaluation on Guessing the Incorrectly Predicted Labels
• Multitask Learning for Class-Imbalanced Discourse Classification
• Semi-supervised Keypoint Localization
• Does Data Augmentation Benefit from Split BatchNorms
• Modeling EEG data distribution with a Wasserstein Generative Adversarial Network to predict RSVP Events
• Automatic Data Augmentation for 3D Medical Image Segmentation
• Low-Resource Multi-Granularity Academic Function Recognition Based on Multiple Prompt Knowledge
• Polyphone Disambiguation in Mandarin Chinese with Semi-Supervised Learning
• Semi-Supervised Crowd Counting from Unlabeled Data
• The Mathematical Foundations of Manifold Learning
• Evolving Neural Selection with Adaptive Regularization
• Text Data Augmentation: Towards better detection of spear-phishing emails
• Adversarially Regularized Policy Learning Guided by Trajectory Optimization
• Identifying Introductions in Podcast Episodes from Automatically Generated Transcripts
• Consistency Regularization for Deep Face Anti-Spoofing
• Substructure Substitution: Structured Data Augmentation for NLP
• Beyond Self-Supervision: A Simple Yet Effective Network Distillation Alternative to Improve Backbones
• Understanding the Generalization Benefit of Model Invariance from a Data Perspective
• Jigsaw Clustering for Unsupervised Visual Representation Learning
• Topological Regularization for Graph Neural Networks Augmentation
• How Out-of-Distribution Data Hurts Semi-Supervised Learning
• Cross-Domain Transfer Learning with CoRTe: Consistent and Reliable Transfer from Black-Box to Lightweight Segmentation Model
• Reducing and Exploiting Data Augmentation Noise through Meta Reweighting Contrastive Learning for Text Classification
• Semi-Supervised Learning with Data Augmentation for End-to-End ASR
• Learning Temporally Invariant and Localizable Features via Data Augmentation for Video Recognition
• Revisiting data augmentation for subspace clustering
• Direct Differentiable Augmentation Search
• Teaching Machine Comprehension with Compositional Explanations
• Reweighting Augmented Samples by Minimizing the Maximal Expected Loss
• UFO${}^2$: A Unified Framework towards Omni-supervised Object Detection
• Learning Speaker Representation with Semi-supervised Learning approach for Speaker Profiling
• Smooth Imitation Learning via Smooth Costs and Smooth Policies
• Semi-supervised learning for medical image classification using imbalanced training data
• Pi-NAS: Improving Neural Architecture Search by Reducing Supernet Training Consistency Shift
• Hypernetwork-Based Augmentation
• Self-Supervised Domain Adaptation with Consistency Training
• Data augmentation with Mobius transformations
• MUSCLE: Strengthening Semi-Supervised Learning Via Concurrent Unsupervised Learning Using Mutual Information Maximization
• Exploiting Large-scale Teacher-Student Training for On-device Acoustic Models
• Multi-view Subword Regularization
• Progressive Representative Labeling for Deep Semi-Supervised Learning
• Learning to Detect Important People in Unlabelled Images for Semi-supervised Important People Detection
• Pseudo-Representation Labeling Semi-Supervised Learning
• Transfer Learning with Pre-trained Conditional Generative Models
• Neural Data-to-Text Generation with LM-based Text Augmentation
• One-bit Supervision for Image Classification: Problem, Solution, and Beyond
• Target Consistency for Domain Adaptation: when Robustness meets Transferability
• Adaptation and Generalization for Unknown Sensitive Factors of Variations
• Webly Supervised Image Classification with Self-Contained Confidence
• Improving Object Detection with Selective Self-supervised Self-training
• Self-Supervised Dynamic Networks for Covariate Shift Robustness
• Social Influence Prediction with Train and Test Time Augmentation for Graph Neural Networks
• Train a One-Million-Way Instance Classifier for Unsupervised Visual Representation Learning
• Enhancing Pre-trained Language Model with Lexical Simplification
• Unsupervised Label Refinement Improves Dataless Text Classification
• Boosting the Performance of Semi-Supervised Learning with Unsupervised Clustering
• AlphaMatch: Improving Consistency for Semi-supervised Learning with Alpha-divergence
• On the Marginal Benefit of Active Learning: Does Self-Supervision Eat Its Cake?
• Learning a Discriminant Latent Space with Neural Discriminant Analysis
• ContraCluster: Learning to Classify without Labels by Contrastive Self-Supervision and Prototype-Based Semi-Supervision
• Vicinal Feature Statistics Augmentation for Federated 3D Medical Volume Segmentation
• On Diversity in Discriminative Neural Networks
• Uncover and Unlearn Nuisances: Agnostic Fully Test-Time Adaptation
• AMPLIFY:Attention-based Mixup for Performance Improvement and Label Smoothing in Transformer
• CGCL: Collaborative Graph Contrastive Learning without Handcrafted Graph Data Augmentations
• Semi-Supervised Vision Transformers
• Boosting Semi-Supervised Face Recognition with Noise Robustness
• Explaining Adversarial Vulnerability with a Data Sparsity Hypothesis
• Reliable Label Bootstrapping for Semi-Supervised Learning
• Gradient-based Data Augmentation for Semi-Supervised Learning
• Meta Label Correction for Noisy Label Learning
• Boosting Unconstrained Face Recognition with Auxiliary Unlabeled Data
• Detecting ESG topics using domain-specific language models and data augmentation approaches
• Matching the Clinical Reality: Accurate OCT-Based Diagnosis From Few Labels
• Regularization with Latent Space Virtual Adversarial Training
• Self-Tuning for Data-Efficient Deep Learning
• Unsupervised Instance Selection with Low-Label, Supervised Learning for Outlier Detection
• Noised Consistency Training for Text Summarization
• Out-of-Manifold Regularization in Contextual Embedding Space for Text Classification
• STraTA: Self-Training with Task Augmentation for Better Few-shot Learning
• Linguistic Knowledge in Data Augmentation for Natural Language Processing: An Example on Chinese Question Matching
• High-Dimensional Bayesian Optimization via Semi-Supervised Learning with Optimized Unlabeled Data Sampling
• Contrastive Forward-Forward: A Training Algorithm of Vision Transformer
• CERM: Context-aware Literature-based Discovery via Sentiment Analysis
• Labeled Data Generation with Inexact Supervision
• Learn to Resolve Conversational Dependency: A Consistency Training Framework for Conversational Question Answering
• Recognising Biomedical Names: Challenges and Solutions
• BiSTF: Bilateral-Branch Self-Training Framework for Semi-Supervised Large-scale Fine-Grained Recognition
• Consistency Training of Multi-exit Architectures for Sensor Data
• Robust Temporal Ensembling for Learning with Noisy Labels
• Efficient data augmentation using graph imputation neural networks
• Semi-Supervised Text Classification via Self-Pretraining
• Revisiting Self-Training for Few-Shot Learning of Language Model
• Scalable Consistency Training for Graph Neural Networks via Self-Ensemble Self-Distillation
• Training Lightweight CNNs for Human-Nanodrone Proximity Interaction from Small Datasets using Background Randomization
• LiDAM: Semi-Supervised Learning with Localized Domain Adaptation and Iterative Matching
• FROST: Faster and more Robust One-shot Semi-supervised Training
• Semi-supervised Sound Event Detection using Random Augmentation and Consistency Regularization
• Semi-supervised Learning of Fetal Anatomy from Ultrasound
• Evaluating Robustness to Input Perturbations for Neural Machine Translation
• Concurrent Training Improves the Performance of Behavioral Cloning from Observation
• Self-supervised learning using consistency regularization of spatio-temporal data augmentation for action recognition
• Langevin Cooling for Domain Translation
• TEST_POSITIVE at W-NUT 2020 Shared Task-3: Joint Event Multi-task Learning for Slot Filling in Noisy Text
• Semi-supervised learning objectives as log-likelihoods in a generative model of data curation