NeurIPS 2025 accepted papers on Label-efficient and robust learning
345 of the 5,812 papers accepted at NeurIPS 2025 (5.9%) are on Label-efficient and robust learning, compared with 7.4% at NeurIPS 2024. They fall into 9 topics.
See NeurIPS 2025 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 · 155 papers
- OrthoLoC: UAV 6-DoF Localization and Calibration Using Orthographic Geodata (4 citations)
- From Experts to a Generalist: Toward General Whole-Body Control for Humanoid Robots (3 citations)
- Mitra: Mixed Synthetic Priors for Enhancing Tabular Foundation Models (3 citations)
- Rethinking Evaluation of Infrared Small Target Detection (3 citations)
- A Standardized Benchmark for Multilabel Antimicrobial Peptide Classification (2 citations)
- BMW: Bidirectionally Memory bank reWriting for Unsupervised Person Re-Identification (2 citations)
- Deferring Concept Bottleneck Models: Learning to Defer Interventions to Inaccurate Experts (2 citations)
- Flatness is Necessary, Neural Collapse is Not: Rethinking Generalization via Grokking (2 citations)
- Harnessing the Universal Geometry of Embeddings (2 citations)
- JADE: Joint Alignment and Deep Embedding for Multi-Slice Spatial Transcriptomics (2 citations)
- On the creation of narrow AI: hierarchy and nonlocality of neural network skills (2 citations)
- Beyond Scalars: Concept-Based Alignment Analysis in Vision Transformers (1 citation)
and 143 more in the interactive map.
Self-supervised and contrastive · 40 papers
- DINO-Foresight: Looking into the Future with DINO (5 citations)
- Joint‑Embedding vs Reconstruction: Provable Benefits of Latent Space Prediction for Self‑Supervised Learning (5 citations)
- Adv-SSL: Adversarial Self-Supervised Representation Learning with Theoretical Guarantees (2 citations)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (2 citations)
- GreenHyperSpectra: A multi-source hyperspectral dataset for global vegetation trait prediction (2 citations)
- UniTraj: Learning a Universal Trajectory Foundation Model from Billion-Scale Worldwide Traces (2 citations)
- FEEL: Quantifying Heterogeneity in Physiological Signals for Generalizable Emotion Recognition (1 citation)
- FocalCodec: Low-Bitrate Speech Coding via Focal Modulation Networks (1 citation)
- Hybrid Autoencoders for Tabular Data: Leveraging Model-Based Augmentation in Low-Label Settings (1 citation)
- Jasmine: Harnessing Diffusion Prior for Self-supervised Depth Estimation (1 citation)
- Adaptive and Multi-scale Affinity Alignment for Hierarchical Contrastive Learning
- CaliGCL: Calibrated Graph Contrastive Learning via Partitioned Similarity and Consistency Discrimination
and 28 more in the interactive map.
Domain adaptation and generalization · 31 papers
- Towards Single-Source Domain Generalized Object Detection via Causal Visual Prompts (2 citations)
- Conformal Inference under High-Dimensional Covariate Shifts via Likelihood-Ratio Regularization (1 citation)
- Domain Adaptive Hashing Retrieval via VLM Assisted Pseudo-Labeling and Dual Space Adaptation (1 citation)
- Generalizable Domain Adaptation for Sim-and-Real Policy Co-Training (1 citation)
- Generalized Category Discovery under Domain Shift: A Frequency Domain Perspective (1 citation)
- Nemotron-CLIMB: Clustering-based Iterative Data Mixture Bootstrapping for Language Model Pre-training (1 citation)
- RrED: Black-box Unsupervised Domain Adaptation via Rectifying-reasoning Errors of Diffusion (1 citation)
- Towards foundational LiDAR world models with efficient latent flow matching (1 citation)
- A Generalized Label Shift Perspective for Cross-Domain Gaze Estimation
- A High-Dimensional Statistical Method for Optimizing Transfer Quantities in Multi-Source Transfer Learning
- Co-Regularization Enhances Knowledge Transfer in High Dimensions
- Continuous Domain Generalization
and 19 more in the interactive map.
OOD and anomaly detection · 28 papers
- ADPretrain: Advancing Industrial Anomaly Detection via Anomaly Representation Pretraining (3 citations)
- Normal-Abnormal Guided Generalist Anomaly Detection (2 citations)
- Revisiting Logit Distributions for Reliable Out-of-Distribution Detection (2 citations)
- An Evidence-Based Post-Hoc Adjustment Framework for Anomaly Detection Under Data Contamination (1 citation)
- Mysteries of the Deep: Role of Intermediate Representations in Out of Distribution Detection (1 citation)
- Out-of-Distribution Generalized Graph Anomaly Detection with Homophily-aware Environment Mixup (1 citation)
- Rethinking Out-of-Distribution Detection and Generalization with Collective Behavior Dynamics (1 citation)
- $\Delta \mathrm{Energy}$: Optimizing Energy Change During Vision-Language Alignment Improves both OOD Detection and OOD Generalization
- Aggregation Hides Out-of-Distribution Generalization Failures from Spurious Correlations
- Anomaly Detection by an Ensemble of Random Pairs of Hyperspheres
- AutoSciDACT: Automated Scientific Discovery through Contrastive Embedding and Hypothesis Testing
- CCL: Causal-aware In-context Learning for Out-of-Distribution Generalization
and 16 more in the interactive map.
Test-time adaptation · 23 papers
- Backpropagation-Free Test-Time Adaptation via Probabilistic Gaussian Alignment (3 citations)
- CLIPTTA: Robust Contrastive Vision-Language Test-Time Adaptation (3 citations)
- Mint: A Simple Test-Time Adaptation of Vision-Language Models against Common Corruptions (2 citations)
- Searching Latent Program Spaces (2 citations)
- Statistics Caching Test-Time Adaptation for Vision-Language Models (2 citations)
- DAA: Amplifying Unknown Discrepancy for Test-Time Discovery (1 citation)
- Generalization vs Specialization under Concept Shift (1 citation)
- Lifelong Test-Time Adaptation via Online Learning in Tracked Low-Dimensional Subspace (1 citation)
- Multi-Expert Distributionally Robust Optimization for Out-of-Distribution Generalization (1 citation)
- PAID: Pairwise Angular-Invariant Decomposition for Continual Test-Time Adaptation (1 citation)
- Buffer layers for Test-Time Adaptation
- D2SA: Dual-Stage Distribution and Slice Adaptation for Efficient Test-Time Adaptation in MRI Reconstruction
and 11 more in the interactive map.
Semi-supervised learning · 21 papers
- RankMatch: A Novel Approach to Semi-Supervised Label Distribution Learning Leveraging Rank Correlation between Labels (3 citations)
- Fourier Clouds: Fast Bias Correction for Imbalanced Semi-Supervised Learning (1 citation)
- Keep It on a Leash: Controllable Pseudo-label Generation Towards Realistic Long-Tailed Semi-Supervised Learning (1 citation)
- OmniGaze: Reward-inspired Generalizable Gaze Estimation in the Wild (1 citation)
- Revisiting Semi-Supervised Learning in the Era of Foundation Models (1 citation)
- Semi-Supervised Regression with Heteroscedastic Pseudo-Labels (1 citation)
- Uncertain Knowledge Graph Completion via Semi-Supervised Confidence Distribution Learning (1 citation)
- VQ-Seg: Vector-Quantized Token Perturbation for Semi-Supervised Medical Image Segmentation (1 citation)
- A Closer Look to Positive-Unlabeled Learning from Fine-grained Perspectives: An Empirical Study
- Evaluating multiple models using labeled and unlabeled data
- Learning from positive and unlabeled examples -Finite size sample bounds
- Lie Detector: Unified Backdoor Detection via Cross-Examination Framework
and 9 more in the interactive map.
Zero-/few-shot and meta-learning · 19 papers
- Repurposing Marigold for Zero-Shot Metric Depth Estimation via Defocus Blur Cues (2 citations)
- Weak-shot Keypoint Estimation via Keyness and Correspondence Transfer (2 citations)
- ChromFound: Towards A Universal Foundation Model for Single-Cell Chromatin Accessibiltiy Data (1 citation)
- DataRater: Meta-Learned Dataset Curation (1 citation)
- Multi-dataset Joint Pre-training of Emotional EEG Enables Generalizable Affective Computing (1 citation)
- TransferTraj: A Vehicle Trajectory Learning Model for Region and Task Transferability (1 citation)
- ZEUS: Zero-shot Embeddings for Unsupervised Separation of Tabular Data (1 citation)
- $\boldsymbol{\lambda}$-Orthogonality Regularization for Compatible Representation Learning
- Boosting Skeleton-based Zero-Shot Action Recognition with Training-Free Test-Time Adaptation
- Dual Alignment Framework for Few-shot Learning with Inter-Set and Intra-Set Shifts
- Enhancing Zero-Shot Black-Box Optimization via Pretrained Models with Efficient Population Modeling, Interaction, and Stable Gradient Approximation
- Evolving and Regularizing Meta-Environment Learner for Fine-Grained Few-Shot Class-Incremental Learning
and 7 more in the interactive map.
Noisy labels · 15 papers
- FlowRefiner: A Robust Traffic Classification Framework against Label Noise (2 citations)
- An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations (1 citation)
- Enhancing Sample Selection Against Label Noise by Cutting Mislabeled Easy Examples (1 citation)
- Noisy Multi-Label Learning through Co-Occurrence-Aware Diffusion (1 citation)
- Confidence-Aware With Prototype Alignment for Partial Multi-label Learning
- Consistency of the $k_n$-nearest neighbor rule under adaptive sampling
- Exploring the Noise Robustness of Online Conformal Prediction
- From Linear to Nonlinear: Provable Weak-to-Strong Generalization through Feature Learning
- From Pretraining to Pathology: How Noise Leads to Catastrophic Inheritance in Medical Models
- Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization
- Online Strategic Classification With Noise and Partial Feedback
- The Effect of Optimal Self-Distillation in Noisy Gaussian Mixture Model
and 3 more in the interactive map.
Knowledge distillation · 13 papers
- Efficient Verified Unlearning For Distillation (1 citation)
- SDPGO: Efficient Self-Distillation Training Meets Proximal Gradient Optimization (1 citation)
- Single-Teacher View Augmentation: Boosting Knowledge Distillation via Angular Diversity (1 citation)
- Synergy Between the Strong and the Weak: Spiking Neural Networks are Inherently Self-Distillers (1 citation)
- Cross-Modal Representational Knowledge Distillation for Enhanced Spike-informed LFP Modeling
- DeepKD: A Deeply Decoupled and Denoised Knowledge Distillation Trainer
- Detecting Data Deviations in Electronic Health Records
- Improving Target Sound Extraction via Disentangled Codec Representations with Privileged Knowledge Distillation
- Learning Task-Agnostic Representations through Multi-Teacher Distillation
- MTL-KD: Multi-Task Learning Via Knowledge Distillation for Generalizable Neural Vehicle Routing Solver
- On the Mechanisms of Weak-to-Strong Generalization: A Theoretical Perspective
- PLD: A Choice-Theoretic List-Wise Knowledge Distillation
and 1 more in the interactive map.