ICML 2018 accepted papers on Label-efficient and robust learning
47 of the 623 papers accepted at ICML 2018 (7.5%) are on Label-efficient and robust learning, compared with 6.2% at ICML 2017. They fall into 7 topics.
See ICML 2018 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 · 27 papers
- Attention-based Deep Multiple Instance Learning (675 citations)
- Disentangling by Factorising (670 citations)
- Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Synthesis (542 citations)
- Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV) (477 citations)
- On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups (260 citations)
- Towards End-to-End Prosody Transfer for Expressive Speech Synthesis with Tacotron (218 citations)
- Hierarchical Multi-Label Classification Networks (138 citations)
- Detecting and Correcting for Label Shift with Black Box Predictors (105 citations)
- Learning to Explain: An Information-Theoretic Perspective on Model Interpretation (105 citations)
- Does Distributionally Robust Supervised Learning Give Robust Classifiers? (104 citations)
- Explicit Inductive Bias for Transfer Learning with Convolutional Networks (100 citations)
- A Semantic Loss Function for Deep Learning with Symbolic Knowledge (70 citations)
and 15 more in the interactive map.
Domain adaptation and generalization · 7 papers
- CyCADA: Cycle-Consistent Adversarial Domain Adaptation (627 citations)
- Learning Semantic Representations for Unsupervised Domain Adaptation (352 citations)
- Augmented CycleGAN: Learning Many-to-Many Mappings from Unpaired Data (329 citations)
- Transfer Learning via Learning to Transfer (124 citations)
- Modeling Sparse Deviations for Compressed Sensing using Generative Models (43 citations)
- JointGAN: Multi-Domain Joint Distribution Learning with Generative Adversarial Nets (24 citations)
- RadialGAN: Leveraging multiple datasets to improve target-specific predictive models using Generative Adversarial Networks (11 citations)
Noisy labels · 5 papers
- MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels (528 citations)
- Dimensionality-Driven Learning with Noisy Labels (240 citations)
- Improving Regression Performance with Distributional Losses (11 citations)
- Active Testing: An Efficient and Robust Framework for Estimating Accuracy (10 citations)
- Training Neural Machines with Trace-Based Supervision
Semi-supervised learning · 3 papers
Knowledge distillation · 2 papers
- Born Again Neural Networks (431 citations)
- Knowledge Transfer with Jacobian Matching (86 citations)
Zero-/few-shot and meta-learning · 2 papers
OOD and anomaly detection · 1 paper
- Deep One-Class Classification (847 citations)
