ICLR 2018 accepted papers on Label-efficient and robust learning
37 of the 337 papers accepted at ICLR 2018 (11.0%) are on Label-efficient and robust learning, compared with 12.1% at ICLR 2017. They fall into 7 topics.
See ICLR 2018 in the interactive map
Submitting to ICLR? 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 · 22 papers
- Active Learning for Convolutional Neural Networks: A Core-Set Approach (678 citations)
- Word translation without parallel data (635 citations)
- On the importance of single directions for generalization (235 citations)
- A Framework for the Quantitative Evaluation of Disentangled Representations (214 citations)
- Learning from Between-class Examples for Deep Sound Recognition (200 citations)
- Unsupervised Machine Translation Using Monolingual Corpora Only (193 citations)
- Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMs (171 citations)
- Evidence Aggregation for Answer Re-Ranking in Open-Domain Question Answering (153 citations)
- All-but-the-Top: Simple and Effective Postprocessing for Word Representations (129 citations)
- Training Confidence-calibrated Classifiers for Detecting Out-of-Distribution Samples (113 citations)
- Learning General Purpose Distributed Sentence Representations via Large Scale Multi-task Learning (99 citations)
- Unsupervised Neural Machine Translation (92 citations)
and 10 more in the interactive map.
Domain adaptation and generalization · 6 papers
- A DIRT-T Approach to Unsupervised Domain Adaptation (431 citations)
- Generalizing Across Domains via Cross-Gradient Training (337 citations)
- Learning to cluster in order to transfer across domains and tasks (119 citations)
- Self-ensembling for visual domain adaptation (60 citations)
- Adversarial Dropout Regularization (50 citations)
- Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation (39 citations)
Semi-supervised learning · 3 papers
Zero-/few-shot and meta-learning · 2 papers
- A Simple Neural Attentive Meta-Learner (875 citations)
- Meta-Learning for Semi-Supervised Few-Shot Classification (726 citations)
Noisy labels · 2 papers
- Synthetic and Natural Noise Both Break Neural Machine Translation (398 citations)
- Learning From Noisy Singly-labeled Data (108 citations)
OOD and anomaly detection · 1 paper
Self-supervised and contrastive · 1 paper
- An efficient framework for learning sentence representations (426 citations)
