AAAI 2021 accepted papers on Label-efficient and robust learning
262 of the 1,654 papers accepted at AAAI 2021 (15.8%) are on Label-efficient and robust learning, compared with 17.1% at AAAI 2020. They fall into 8 topics.
See AAAI 2021 in the interactive map
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 · 129 papers
- Contrastive Clustering (648 citations)
- Document-Level Relation Extraction with Adaptive Thresholding and Localized Context Pooling (309 citations)
- Robust Lightweight Facial Expression Recognition Network with Label Distribution Training (283 citations)
- A Joint Training Dual-MRC Framework for Aspect Based Sentiment Analysis (215 citations)
- Bidirectional Machine Reading Comprehension for Aspect Sentiment Triplet Extraction (208 citations)
- Camera-Aware Proxies for Unsupervised Person Re-Identification (208 citations)
- Exploiting Sample Uncertainty for Domain Adaptive Person Re-Identification (176 citations)
- To Choose or to Fuse? Scale Selection for Crowd Counting (174 citations)
- Bag of Tricks for Long-Tailed Visual Recognition with Deep Convolutional Neural Networks (131 citations)
- Pre-training Context and Time Aware Location Embeddings from Spatial-Temporal Trajectories for User Next Location Prediction (122 citations)
- Generalize a Small Pre-trained Model to Arbitrarily Large TSP Instances (120 citations)
- Fairness-aware News Recommendation with Decomposed Adversarial Learning (119 citations)
and 117 more in the interactive map.
Domain adaptation and generalization · 39 papers
- Domain General Face Forgery Detection by Learning to Weight (152 citations)
- Embracing Domain Differences in Fake News: Cross-domain Fake News Detection using Multi-modal Data (152 citations)
- Plug-and-Play Domain Adaptation for Cross-Subject EEG-based Emotion Recognition (146 citations)
- A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Data (144 citations)
- Bi-Classifier Determinacy Maximization for Unsupervised Domain Adaptation (126 citations)
- Generalizable Representation Learning for Mixture Domain Face Anti-Spoofing (125 citations)
- CrossNER: Evaluating Cross-Domain Named Entity Recognition (121 citations)
- Self-Domain Adaptation for Face Anti-Spoofing (104 citations)
- DAST: Unsupervised Domain Adaptation in Semantic Segmentation Based on Discriminator Attention and Self-Training (84 citations)
- Latent Independent Excitation for Generalizable Sensor-based Cross-Person Activity Recognition (83 citations)
- Unsupervised Model Adaptation for Continual Semantic Segmentation (49 citations)
- How Does the Combined Risk Affect the Performance of Unsupervised Domain Adaptation Approaches? (46 citations)
and 27 more in the interactive map.
Zero-/few-shot and meta-learning · 22 papers
- Learning a Few-shot Embedding Model with Contrastive Learning (182 citations)
- Partial Is Better Than All: Revisiting Fine-tuning Strategy for Few-shot Learning (153 citations)
- Generalized Zero-Shot Learning via Disentangled Representation (78 citations)
- Few-shot Learning for Multi-label Intent Detection (58 citations)
- Learning Intact Features by Erasing-Inpainting for Few-shot Classification (58 citations)
- Looking Wider for Better Adaptive Representation in Few-Shot Learning (43 citations)
- Attributes-Guided and Pure-Visual Attention Alignment for Few-Shot Recognition (42 citations)
- Few-Shot One-Class Classification via Meta-Learning (41 citations)
- How Linguistically Fair Are Multilingual Pre-Trained Language Models? (41 citations)
- ‘Less Than One’-Shot Learning: Learning N Classes From M < N Samples (39 citations)
- Task Aligned Generative Meta-learning for Zero-shot Learning (37 citations)
- Proxy Synthesis: Learning with Synthetic Classes for Deep Metric Learning (34 citations)
and 10 more in the interactive map.
Self-supervised and contrastive · 19 papers
- TabNet: Attentive Interpretable Tabular Learning (1,769 citations)
- Self-Supervised Hypergraph Convolutional Networks for Session-based Recommendation (574 citations)
- HR-Depth: High Resolution Self-Supervised Monocular Depth Estimation (300 citations)
- Self-supervised Multi-view Stereo via Effective Co-Segmentation and Data-Augmentation (79 citations)
- Learning an Effective Context-Response Matching Model with Self-Supervised Tasks for Retrieval-based Dialogues (49 citations)
- Coupling Macro-Sector-Micro Financial Indicators for Learning Stock Representations with Less Uncertainty (39 citations)
- A Simple and Effective Self-Supervised Contrastive Learning Framework for Aspect Detection (37 citations)
- Visualization of Supervised and Self-Supervised Neural Networks via Attribution Guided Factorization (32 citations)
- Deep Metric Learning with Self-Supervised Ranking (24 citations)
- Context Matters: Graph-based Self-supervised Representation Learning for Medical Images (21 citations)
- Exploiting Behavioral Consistence for Universal User Representation (21 citations)
- Interpretable Self-Supervised Facial Micro-Expression Learning to Predict Cognitive State and Neurological Disorders (15 citations)
and 7 more in the interactive map.
Semi-supervised learning · 19 papers
- Semi-supervised Medical Image Segmentation through Dual-task Consistency (689 citations)
- Curriculum Labeling: Revisiting Pseudo-Labeling for Semi-Supervised Learning (285 citations)
- What the Role is vs. What Plays the Role: Semi-Supervised Event Argument Extraction via Dual Question Answering (66 citations)
- PTN: A Poisson Transfer Network for Semi-supervised Few-shot Learning (24 citations)
- SALNet: Semi-supervised Few-Shot Text Classification with Attention-based Lexicon Construction (24 citations)
- Inferring Emotion from Large-scale Internet Voice Data: A Semi-supervised Curriculum Augmentation based Deep Learning Approach (22 citations)
- Revisiting Iterative Back-Translation from the Perspective of Compositional Generalization (21 citations)
- Explanation Consistency Training: Facilitating Consistency-Based Semi-Supervised Learning with Interpretability (20 citations)
- SHOT-VAE: Semi-supervised Deep Generative Models With Label-aware ELBO Approximations (18 citations)
- PASSLEAF: A Pool-bAsed Semi-Supervised LEArning Framework for Uncertain Knowledge Graph Embedding (17 citations)
- Weakly-Supervised Hierarchical Models for Predicting Persuasive Strategies in Good-faith Textual Requests (15 citations)
- Hierarchical Information Passing Based Noise-Tolerant Hybrid Learning for Semi-Supervised Human Parsing (14 citations)
and 7 more in the interactive map.
Knowledge distillation · 17 papers
- Cross-Layer Distillation with Semantic Calibration (311 citations)
- Show, Attend and Distill: Knowledge Distillation via Attention-based Feature Matching (153 citations)
- Peer Collaborative Learning for Online Knowledge Distillation (133 citations)
- Reinforced Multi-Teacher Selection for Knowledge Distillation (126 citations)
- Progressive Network Grafting for Few-Shot Knowledge Distillation (47 citations)
- Diverse Knowledge Distillation for End-to-End Person Search (39 citations)
- Data-Free Knowledge Distillation with Soft Targeted Transfer Set Synthesis (29 citations)
- Improving Tree-Structured Decoder Training for Code Generation via Mutual Learning (24 citations)
- Stochastic Precision Ensemble: Self-Knowledge Distillation for Quantized Deep Neural Networks (21 citations)
- PSSM-Distil: Protein Secondary Structure Prediction (PSSP) on Low-Quality PSSM by Knowledge Distillation with Contrastive Learning (16 citations)
- Flexible Non-Autoregressive Extractive Summarization with Threshold: How to Extract a Non-Fixed Number of Summary Sentences (13 citations)
- Knowledge Refinery: Learning from Decoupled Label (12 citations)
and 5 more in the interactive map.
Noisy labels · 13 papers
- Meta Label Correction for Noisy Label Learning (199 citations)
- Beyond Class-Conditional Assumption: A Primary Attempt to Combat Instance-Dependent Label Noise (112 citations)
- Label Confusion Learning to Enhance Text Classification Models (67 citations)
- Learning to Purify Noisy Labels via Meta Soft Label Corrector (66 citations)
- Learning from Crowds by Modeling Common Confusions (37 citations)
- Robustness of Accuracy Metric and its Inspirations in Learning with Noisy Labels (26 citations)
- Learning from Noisy Labels with Complementary Loss Functions (23 citations)
- Tackling Instance-Dependent Label Noise via a Universal Probabilistic Model (23 citations)
- Empower Distantly Supervised Relation Extraction with Collaborative Adversarial Training (19 citations)
- Analysing the Noise Model Error for Realistic Noisy Label Data (11 citations)
- Denoising Distantly Supervised Named Entity Recognition via a Hypergeometric Probabilistic Model (10 citations)
- Boosting Multi-task Learning Through Combination of Task Labels – with Applications in ECG Phenotyping (7 citations)
and 1 more in the interactive map.
OOD and anomaly detection · 4 papers
- DecAug: Out-of-Distribution Generalization via Decomposed Feature Representation and Semantic Augmentation (60 citations)
- Deep Verifier Networks: Verification of Deep Discriminative Models with Deep Generative Models (33 citations)
- MASKER: Masked Keyword Regularization for Reliable Text Classification (28 citations)
- Multidimensional Uncertainty-Aware Evidential Neural Networks (22 citations)
