ICLR 2026 accepted papers on Label-efficient and robust learning
259 of the 5,358 papers accepted at ICLR 2026 (4.8%) are on Label-efficient and robust learning, compared with 6.2% at ICLR 2025. They fall into 9 topics.
See ICLR 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Recognition and long-tail · 98 papers
- On the Theoretical Limitations of Embedding-Based Retrieval (6 citations)
- DeepFRC: An End-to-End Deep Learning Model for Functional Registration and Classification (2 citations)
- AnyUp: Universal Feature Upsampling (1 citation)
- Boosting Open Set Recognition Performance through Modulated Representation Learning (1 citation)
- Omni-iEEG: A Large-Scale, Comprehensive iEEG Dataset and Benchmark for Epilepsy Research (1 citation)
- Adaptive Regularization for Large-Scale Sparse Feature Embedding Models
- Behavioral Embeddings of Programs: A Quasi-Dynamic Approach for Optimization Prediction
- Beyond Hearing: Learning Task-agnostic ExG Representations from Earphones via Physiology-informed Tokenization
- Characterizing Human Semantic Navigation in Concept Production as Trajectories in Embedding Space
- Characterizing Pattern Matching and Its Limits on Compositional Task Structures
- Concepts' Information Bottleneck Models
- Conformal Prediction for Long-Tailed Classification
and 86 more in the interactive map.
Self-supervised and contrastive · 37 papers
- One protein is all you need (2 citations)
- CryoLVM: Self-supervised Learning from Cryo-EM Density Maps with Large Vision Models (1 citation)
- FM4NPP: A Scaling Foundation Model for Nuclear and Particle Physics (1 citation)
- TerraFM: A Scalable Foundation Model for Unified Multisensor Earth Observation (1 citation)
- Accelerating Benchmarking of Functional Connectivity Modeling via Structure-aware Core-set Selection
- Animal behavioral analysis and neural encoding with transformer-based self-supervised pretraining
- CARL: Camera-Agnostic Representation Learning for Spectral Image Analysis
- Difficult Examples Hurt Unsupervised Contrastive Learning: A Theoretical Perspective
- Equivariant Splitting: Self-supervised learning from incomplete data
- Maximizing Incremental Information Entropy for Contrastive Learning
- MULTIMODALITY AS SUPERVISION: SELF-SUPERVISED SPECIALIZATION TO THE TEST ENVIRONMENT VIA MULTIMODALITY
- On the Alignment Between Supervised and Self-Supervised Contrastive Learning
and 25 more in the interactive map.
Knowledge distillation · 32 papers
- Beyond Student: An Asymmetric Network for Neural Network Inheritance
- BIRD: Behavior Induction via Representation-structure Distillation
- Boomerang Distillation Enables Zero-Shot Model Size Interpolation
- Cross-Tokenizer Likelihood Scoring Algorithms for Language Model Distillation
- Dataset Distillation for Memorized Data: Soft Labels can Leak Held-Out Teacher Knowledge
- DUET: DISTILLED LLM UNLEARNING FROM AN EFFICIENTLY CONTEXTUALIZED TEACHER
- Explain in Your Own Words: Improving Reasoning via Token-Selective Dual Knowledge Distillation
- Exploring Knowledge Purification in Multi-Teacher Knowledge Distillation for LLMs
- Generative Diffusion Prior Distillation for Long-Context Knowledge Transfer
- Grounding and Enhancing Informativeness and Utility in Dataset Distillation
- In Good GRACES: Principled Teacher Selection for Knowledge Distillation
- Learning Shrinks the Hard Tail: Training‑Dependent Inference Scaling in a Solvable Linear Model
and 20 more in the interactive map.
Test-time adaptation · 25 papers
- NEO — No-Optimization Test-Time Adaptation through Latent Re-Centering (1 citation)
- Adaptive Debiasing Tsallis Entropy for Test-Time Adaptation
- Adaptive Test-Time Training for Predicting Need for Invasive Mechanical Ventilation in Multi-Center Cohorts
- Architecture-Agnostic Test-Time Adaptation via Backprop-Free Embedding Alignment
- Flatness Guided Test-Time Adaptation for Vision-Language Models
- Get RICH or Die Scaling: Profitably Trading Inference Compute for Robustness
- IMSE: Intrinsic Mixture of Spectral Experts Fine-tuning for Test-Time Adaptation
- Mitigating Spurious Correlation via Distributionally Robust Learning with Hierarchical Ambiguity Sets
- PriorGuide: Test-Time Prior Adaptation for Simulation-Based Inference
- Robust Test-time Video-Text Retrieval: Benchmarking and Adapting for Query Shifts
- Test-Time Mixture of World Models for Embodied Agents in Dynamic Environments
- When and Where to Reset Matters for Long-Term Test-Time Adaptation
and 13 more in the interactive map.
OOD and anomaly detection · 22 papers
- AP-OOD: Attention Pooling for Out-of- Distribution Detection
- Boosting for Predictive Sufficiency
- Dual Distillation for Few-Shot Anomaly Detection
- Foundation Visual Encoders Are Secretly Few-Shot Anomaly Detectors
- GradPCA: Leveraging NTK Alignment for Reliable Out-of-Distribution Detection
- LAMDA: A Longitudinal Android Malware Benchmark for Concept Drift Analysis
- MRAD: Zero-Shot Anomaly Detection with Memory-Driven Retrieval
- PSDNorm: Temporal Normalization for Deep Learning in Sleep Staging
- ReTabAD: A Benchmark for Restoring Semantic Context in Tabular Anomaly Detection
- SCOPED: Score–Curvature Out-of-distribution Proximity Evaluator for Diffusion
- Towards a Certificate of Trust: Task-Aware OOD Detection for Scientific AI
- UniOD: A Universal Model for Outlier Detection across Diverse Domains
and 10 more in the interactive map.
Domain adaptation and generalization · 16 papers
- Cross-Domain Policy Optimization via Bellman Consistency and Hybrid Critics
- ELViS: Efficient Visual Similarity from Local Descriptors that Generalizes Across Domains
- HEEGNet: Hyperbolic Embeddings for EEG
- PAS: Estimating the target accuracy before domain adaptation
- Residual Feature Integration is Sufficient to Prevent Negative Transfer
- Transfer Learning in Infinite Width Feature Learning Networks
- Universal Multi-Domain Translation via Diffusion Routers
- Conformalized Hierarchical Calibration for Uncertainty-Aware Adaptive Hashing
- Distributionally Robust Classification for Multi-source Unsupervised Domain Adaptation
- Exploring Mode Connectivity in Krylov Subspace for Domain Generalization
- Minimax-Optimal Aggregation for Density Ratio Estimation
- Personalized Feature Translation for Expression Recognition: An Efficient Source-Free Domain Adaptation Method
and 4 more in the interactive map.
Zero-/few-shot and meta-learning · 14 papers
- Relational Transformer: Toward Zero-Shot Foundation Models for Relational Data (1 citation)
- Distilling and Adapting: A Topology-Aware Framework for Zero-Shot Interaction Prediction in Multiplex Biological Networks
- Evaluating Data Influence in Meta Learning
- Exploiting Low-Dimensional Manifold of Features for Few-shot Whole Slide Image Classification
- LiteGuard: Efficient Task-Agnostic Model Fingerprinting with Enhanced Generalization
- Self-Predictive Representations for Combinatorial Generalization in Behavioral Cloning
- SubDyve: Subgraph-Driven Dynamic Propagation for Virtual Screening Enhancement Controlling False Positive
- Zero-Shot Adaptation of Behavioral Foundation Models to Unseen Dynamics
- $\mu$LO: Compute-Efficient Meta-Generalization of Learned Optimizers
- BTZSC: A Benchmark for Zero-Shot Text Classification Across Cross-Encoders, Embedding Models, and Rerankers
- Decoupling Primitive with Experts: Dynamic Feature Alignment for Compositional Zero-Shot Learning
- Instance-wise Adaptive Scheduling via Derivative-Free Meta-Learning
and 2 more in the interactive map.
Noisy labels · 10 papers
- Noise-Aware Generalization: Robustness to In-Domain Noise and Out-of-Domain Generalization (1 citation)
- Noisy-Pair Robust Representation Alignment for Positive-Unlabeled Learning
- Random Label Prediction Heads for Studying Memorization in Deep Neural Networks
- Resurfacing the Instance-only Dependent Label Noise Model through Loss Correction
- Trapped by simplicity: When Transformers fail to learn from noisy features
- Distributional Consistency Loss: Beyond Pointwise Data Terms in Inverse Problems
- Enhancing Learning with Noisy Labels via Rockafellian Relaxation
- LANE: Label-Aware Noise Elimination for Fine-Grained Text Classification
- Towards a Foundation Model for Crowdsourced Label Aggregation
- TrainRef: Curating Data with Label Distribution and Minimal Reference for Accurate Prediction and Reliable Confidence
Semi-supervised learning · 5 papers
- In Context Semi-Supervised Learning
- Learning Dynamics of Logits Debiasing for Long-Tailed Semi-Supervised Learning
- Adaptive Conformal Guidance for Learning under Uncertainty
- CoLA: Co-Calibrated Logit Adjustment for Long-Tailed Semi-Supervised Learning
- SCAD: Super-Class-Aware Debiasing for Long-Tailed Semi-Supervised Learning