ICML 2019 accepted papers on Label-efficient and robust learning
65 of the 774 papers accepted at ICML 2019 (8.4%) are on Label-efficient and robust learning, compared with 7.5% at ICML 2018. They fall into 8 topics.
See ICML 2019 in the interactive map
Submitting to ICML? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Recognition and long-tail · 35 papers
- Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations (767 citations)
- Do ImageNet Classifiers Generalize to ImageNet? (541 citations)
- Similarity of Neural Network Representations Revisited (513 citations)
- A Theoretical Analysis of Contrastive Unsupervised Representation Learning (198 citations)
- What is the Effect of Importance Weighting in Deep Learning? (112 citations)
- Understanding the Origins of Bias in Word Embeddings (106 citations)
- MeanSum: A Neural Model for Unsupervised Multi-Document Abstractive Summarization (99 citations)
- Concrete Autoencoders: Differentiable Feature Selection and Reconstruction (86 citations)
- Almost Unsupervised Text to Speech and Automatic Speech Recognition (79 citations)
- Unsupervised Deep Learning by Neighbourhood Discovery (64 citations)
- SelectiveNet: A Deep Neural Network with an Integrated Reject Option (60 citations)
- Addressing the Loss-Metric Mismatch with Adaptive Loss Alignment (48 citations)
and 23 more in the interactive map.
Domain adaptation and generalization · 10 papers
- Bridging Theory and Algorithm for Domain Adaptation (326 citations)
- Transferability vs. Discriminability: Batch Spectral Penalization for Adversarial Domain Adaptation (268 citations)
- On Learning Invariant Representations for Domain Adaptation (179 citations)
- Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers (171 citations)
- Domain Agnostic Learning with Disentangled Representations (150 citations)
- Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment (76 citations)
- Feature-Critic Networks for Heterogeneous Domain Generalization (72 citations)
- Learning What and Where to Transfer (24 citations)
- Multivariate-Information Adversarial Ensemble for Scalable Joint Distribution Matching (3 citations)
- Towards Accurate Model Selection in Deep Unsupervised Domain Adaptation
Noisy labels · 8 papers
- Unsupervised Label Noise Modeling and Loss Correction (193 citations)
- SELFIE: Refurbishing Unclean Samples for Robust Deep Learning (190 citations)
- How does Disagreement Help Generalization against Label Corruption? (151 citations)
- Noise2Self: Blind Denoising by Self-Supervision (117 citations)
- Combating Label Noise in Deep Learning using Abstention (103 citations)
- Understanding and Utilizing Deep Neural Networks Trained with Noisy Labels (91 citations)
- Robust Inference via Generative Classifiers for Handling Noisy Labels (56 citations)
- Exploiting Worker Correlation for Label Aggregation in Crowdsourcing (36 citations)
Zero-/few-shot and meta-learning · 6 papers
- AutoVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss (255 citations)
- Hierarchically Structured Meta-learning (126 citations)
- TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning (76 citations)
- LGM-Net: Learning to Generate Matching Networks for Few-Shot Learning (50 citations)
- Co-Representation Network for Generalized Zero-Shot Learning (46 citations)
- Learning Classifiers for Target Domain with Limited or No Labels (8 citations)
Semi-supervised learning · 3 papers
OOD and anomaly detection · 1 paper
- Anomaly Detection With Multiple-Hypotheses Predictions (36 citations)
Test-time adaptation · 1 paper
- Fast Context Adaptation via Meta-Learning (177 citations)
Knowledge distillation · 1 paper
- Zero-Shot Knowledge Distillation in Deep Networks (82 citations)
