Swin Transformer reaches 87.3% ImageNet accuracy and sets new records on COCO detection and ADE20K segmentation by replacing global self-attention with shifted-window local attention inside a hierarchical pyramid.
hub Mixed citations
mixup: Beyond Empirical Risk Minimization
Mixed citation behavior. Most common role is background (47%).
90
Pith papers citing it
4,767
external citations · Pith
Background
47% of classified citations
abstract
Large deep neural networks are powerful, but exhibit undesirable behaviors such as memorization and sensitivity to adversarial examples. In this work, we propose mixup, a simple learning principle to alleviate these issues. In essence, mixup trains a neural network on convex combinations of pairs of examples and their labels. By doing so, mixup regularizes the neural network to favor simple linear behavior in-between training examples. Our experiments on the ImageNet-2012, CIFAR-10, CIFAR-100, Google commands and UCI datasets show that mixup improves the generalization of state-of-the-art neural network architectures. We also find that mixup reduces the memorization of corrupt labels, increases the robustness to adversarial examples, and stabilizes the training of generative adversarial networks.
hub tools
citation-role summary
background 6
method 6
baseline 2
other 1
citation-polarity summary
claims ledger
- abstract Large deep neural networks are powerful, but exhibit undesirable behaviors such as memorization and sensitivity to adversarial examples. In this work, we propose mixup, a simple learning principle to alleviate these issues. In essence, mixup trains a neural network on convex combinations of pairs of examples and their labels. By doing so, mixup regularizes the neural network to favor simple linear behavior in-between training examples. Our experiments on the ImageNet-2012, CIFAR-10, CIFAR-100, Google commands and UCI datasets show that mixup improves the generalization of state-of-the-art neur
co-cited works
representative citing papers
TCP-SSM conditions stable poles on visual tokens to explicitly control memory decay and oscillation in SSMs, cutting computation up to 44% while matching or exceeding accuracy on classification, segmentation, and detection.
An efficiently computable HS-Jacobian acts as a conservative mapping for projections onto polyhedral sets, supporting provably convergent Adam-based end-to-end training of linearly constrained deep neural networks.
LG-CoTrain, an LLM-guided co-training method, outperforms classical semi-supervised baselines for crisis tweet classification in low-resource settings with 5-25 labeled examples per class.
LookWhen factorizes video recognition into learning when, where, and what to compute via uniqueness-based token selection and dual-teacher distillation, achieving better accuracy-FLOPs trade-offs than baselines on multiple datasets.
PARSE improves domain generalization accuracy by factoring recognition into visual primitives and their spatial relational compositions learned end-to-end with differentiable predicates.
SignMAE uses segmentation-driven masking in a mask-and-reconstruct self-supervised task to learn fine-grained sign representations, achieving state-of-the-art accuracy on WLASL, NMFs-CSL, and Slovo with fewer frames and modalities.
A replay method for continual face forgery detection condenses real-fake distribution discrepancies into compact maps and synthesizes compatible samples from current real faces to reduce forgetting under tight memory budgets without storing historical images.
Machine unlearning conflates reversing the influence of specific training examples (untraining) with removing the full underlying distribution or behavior (unlearning).
DREAM introduces Masking Warmup and Semantically Aligned Decoding to let a single encoder handle both contrastive alignment and masked generation, yielding gains over CLIP and FLUID on understanding and generation benchmarks.
Chronos pretrains transformer models on tokenized time series to deliver strong zero-shot forecasting across diverse domains.
The DFDC dataset is the largest public collection of face-swapped videos and supports detectors that generalize to in-the-wild deepfakes.
Lightweight CNN with separable convolutions, hierarchical augmentation and power-based label smoothing reaches 84% cross-environment beam prediction accuracy on two real DeepSense 6G scenarios while cutting parameters by 52x and complexity by 79x versus ResNet.
Evaluates four distribution shifts in sensor-based HAR, finds diversity shifts dominate, and shows 28 DG methods only marginally beat ERM while releasing open benchmarks.
A periodic multi-excitation level procedure leverages system non-linearities to generate images for pre-trained CNNs in vibration-based intelligent fault diagnosis under data scarcity, validated experimentally on a railway pantograph.
Tri-Info uses three information theory signals on action diversity, temporal consistency, and state coupling to predict VLA model failures with cross-domain generalization to 83% real-world accuracy.
Unsupervised symmetry discovery via shallow group-convolutional networks recovers latent domains from linear measurements of random fields by learning symmetry actions under stationarity and locality constraints.
MindAlign decodes inner speech from fMRI via subject-specific neural-semantic alignment into a multimodal space followed by prompting of a frozen LM, outperforming baselines and generalizing across subjects.
SNR-ST-Mix is a geometry- and expression-aware mixup augmentation that constrains interpolation to k-nearest spatial neighbors and weights by transcriptomic similarity for spatial transcriptomics imputation.
A LoRA-adapted diffusion model generates surveillance-style pedestrian images whose labels are verified by a Bayesian classifier on BLIP scores, improving PAR accuracy over naive synthetic labeling.
A causality-inspired FedDG framework with device style intervention network, counterfactual text augmentation, and gradient alignment outperforms baselines on leave-one-device-out validation for RSC on ICBHI and SPRSound datasets.
Representation-conditioned diffusion models generate synthetic ImageNet data that trains classifiers to higher top-1 accuracy than class-conditioned generation (+10.76 pp) or real data (+2.0 pp when scaled).
GAMR introduces geometric-aware manifold regularization via virtual outlier synthesis to enhance intra-class compactness and inter-class separation, improving robustness to noisy labels beyond passive sample filtering.
HamBR uses Spherical HMC to probe ambiguous regions and synthesize virtual outliers with energy-based repulsion to restore decision boundaries degraded by noisy labels, achieving SOTA on CIFAR and real-world benchmarks.
citing papers explorer
-
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
Swin Transformer reaches 87.3% ImageNet accuracy and sets new records on COCO detection and ADE20K segmentation by replacing global self-attention with shifted-window local attention inside a hierarchical pyramid.
-
TCP-SSM: Efficient Vision State Space Models with Token-Conditioned Poles
TCP-SSM conditions stable poles on visual tokens to explicitly control memory decay and oscillation in SSMs, cutting computation up to 44% while matching or exceeding accuracy on classification, segmentation, and detection.
-
Efficient and provably convergent end-to-end training of deep neural networks with linear constraints
An efficiently computable HS-Jacobian acts as a conservative mapping for projections onto polyhedral sets, supporting provably convergent Adam-based end-to-end training of linearly constrained deep neural networks.
-
LLM-guided Semi-Supervised Approaches for Social Media Crisis Data Classification
LG-CoTrain, an LLM-guided co-training method, outperforms classical semi-supervised baselines for crisis tweet classification in low-resource settings with 5-25 labeled examples per class.
-
LookWhen? Fast Video Recognition by Learning When, Where, and What to Compute
LookWhen factorizes video recognition into learning when, where, and what to compute via uniqueness-based token selection and dual-teacher distillation, achieving better accuracy-FLOPs trade-offs than baselines on multiple datasets.
-
Domain Generalization through Spatial Relation Induction over Visual Primitives
PARSE improves domain generalization accuracy by factoring recognition into visual primitives and their spatial relational compositions learned end-to-end with differentiable predicates.
-
SignMAE: Segmentation-Driven Self-Supervised Learning for Sign Language Recognition
SignMAE uses segmentation-driven masking in a mask-and-reconstruct self-supervised task to learn fine-grained sign representations, achieving state-of-the-art accuracy on WLASL, NMFs-CSL, and Slovo with fewer frames and modalities.
-
Direct Discrepancy Replay: Distribution-Discrepancy Condensation and Manifold-Consistent Replay for Continual Face Forgery Detection
A replay method for continual face forgery detection condenses real-fake distribution discrepancies into compact maps and synthesizes compatible samples from current real faces to reduce forgetting under tight memory budgets without storing historical images.
-
Is your algorithm unlearning or untraining?
Machine unlearning conflates reversing the influence of specific training examples (untraining) with removing the full underlying distribution or behavior (unlearning).
-
Unifying Contrastive and Generative Objectives for Visual Understanding and Text-to-Image Generation
DREAM introduces Masking Warmup and Semantically Aligned Decoding to let a single encoder handle both contrastive alignment and masked generation, yielding gains over CLIP and FLUID on understanding and generation benchmarks.
-
Chronos: Learning the Language of Time Series
Chronos pretrains transformer models on tokenized time series to deliver strong zero-shot forecasting across diverse domains.
-
The DeepFake Detection Challenge (DFDC) Dataset
The DFDC dataset is the largest public collection of face-swapped videos and supports detectors that generalize to in-the-wild deepfakes.
-
Lightweight Vision-Aided Beam Tracking for Cross-Environment mmWave Communications
Lightweight CNN with separable convolutions, hierarchical augmentation and power-based label smoothing reaches 84% cross-environment beam prediction accuracy on two real DeepSense 6G scenarios while cutting parameters by 52x and complexity by 79x versus ResNet.
-
Assessing Distribution Shift in Human Activity Recognition for Domain Generalization
Evaluates four distribution shifts in sensor-based HAR, finds diversity shifts dominate, and shows 28 DG methods only marginally beat ERM while releasing open benchmarks.
-
Leveraging systems' non-linearity to tackle the scarcity of data in the design of Intelligent Fault Diagnosis Systems
A periodic multi-excitation level procedure leverages system non-linearities to generate images for pre-trained CNNs in vibration-based intelligent fault diagnosis under data scarcity, validated experimentally on a railway pantograph.
-
Tri-Info: Generalizable, Interpretable Failure Prediction for VLA Models via Information Theory
Tri-Info uses three information theory signals on action diversity, temporal consistency, and state coupling to predict VLA model failures with cross-domain generalization to 83% real-world accuracy.
-
Blind Recovery of Latent Domains via Unsupervised Symmetry Discovery
Unsupervised symmetry discovery via shallow group-convolutional networks recovers latent domains from linear measurements of random fields by learning symmetry actions under stationarity and locality constraints.
-
MindAlign: Decoding Inner Speech from fMRI Signals via Multimodal Embedding Alignment under Limited Data
MindAlign decodes inner speech from fMRI via subject-specific neural-semantic alignment into a multimodal space followed by prompting of a frozen LM, outperforming baselines and generalizing across subjects.
-
SNR-ST-Mix: Sample-specific Neighborhood Regression Mixup for Augmented Spatial Transcriptomics Imputation with Deep Neural Network
SNR-ST-Mix is a geometry- and expression-aware mixup augmentation that constrains interpolation to k-nearest spatial neighbors and weights by transcriptomic similarity for spatial transcriptomics imputation.
-
ReSAGE-PAR: Representational Similarity Assessment for Generative Expansion in Pedestrian Attribute Recognition
A LoRA-adapted diffusion model generates surveillance-style pedestrian images whose labels are verified by a Bayesian classifier on BLIP scores, improving PAR accuracy over naive synthetic labeling.
-
Mitigating Stethoscope-Induced Shortcuts in Respiratory Sound Classification under Federated Domain Generalization with Causality-Inspired Interventions
A causality-inspired FedDG framework with device style intervention network, counterfactual text augmentation, and gradient alignment outperforms baselines on leave-one-device-out validation for RSC on ICBHI and SPRSound datasets.
-
Representation-Conditioned Diffusion Models for Guided Training Data Generation
Representation-conditioned diffusion models generate synthetic ImageNet data that trains classifiers to higher top-1 accuracy than class-conditioned generation (+10.76 pp) or real data (+2.0 pp when scaled).
-
GAMR: Geometric-Aware Manifold Regularization with Virtual Outlier Synthesis for Learning with Noisy Labels
GAMR introduces geometric-aware manifold regularization via virtual outlier synthesis to enhance intra-class compactness and inter-class separation, improving robustness to noisy labels beyond passive sample filtering.
-
HamBR: Active Decision Boundary Restoration Based on Hamiltonian Dynamics for Learning with Noisy Labels
HamBR uses Spherical HMC to probe ambiguous regions and synthesize virtual outliers with energy-based repulsion to restore decision boundaries degraded by noisy labels, achieving SOTA on CIFAR and real-world benchmarks.
-
LiBaGS: Lightweight Boundary Gap Synthesis for Targeted Synthetic Data Selection
LiBaGS scores and selects synthetic data near decision boundaries using proximity, uncertainty, density, and validity, with boundary-gap allocation and marginal stopping to improve training accuracy.
-
Cross-Sample Relational Fusion: Unifying Domain Generalization and Class-Incremental Learning
CORF unifies domain generalization and class-incremental learning via selective sample refinement with spatial maps and confidence weighting plus cascaded relational distillation.
-
ICDAR 2026 Competition on Writer Identification and Pen Classification from Hand-Drawn Circles
CircleID introduces a controlled dataset of 46,155 circles from 66 writers and 8 pens, with competition results showing top accuracies of 64.8% for open-set writer identification and 92.7% for pen classification.
-
Leveraging Image Generators to Address Data Scarcity: The Gen4Regen Dataset for Forest Regeneration Mapping
Prompt-generated image-mask pairs, mixed with real UAV imagery at a 40:60 ratio, lift forest-regeneration segmentation by >15 F1 points over supervised baselines and sharply improve rare-species F1.
-
LEGO: LoRA-Enabled Generator-Oriented Framework for Synthetic Image Detection
A detector that trains a separate LoRA adapter per generator and mixes them with a learned router reports the best mean accuracy on AIGIBench and Chameleon, with weaker per-dataset results on several subsets.
-
Cheeger--Hodge Contrastive Learning for Structurally Robust Graph Representation Learning
CHCL aligns a Cheeger-Hodge joint signature across graph augmentations to produce embeddings that remain stable under local structural changes.
-
Beyond Binary Contrast: Modeling Continuous Skeleton Action Spaces with Transitional Anchors
TranCLR models continuous skeleton action spaces with transitional anchors and multi-level manifold calibration, yielding smoother and more accurate representations than binary contrastive methods.
-
PAC-Bayes Bounds for Gibbs Posteriors via Singular Learning Theory
PAC-Bayes bounds for Gibbs posteriors are obtained via singular learning theory, producing explicit and tighter posterior-averaged risk bounds that adapt to data structure in overparameterized models.
-
Human Gaze-based Dual Teacher Guidance Learning for Semi-Supervised Medical Image Segmentation
HG-DTGL integrates human gaze as an extra teacher in mean-teacher learning via GazeMix, MGP module and Gaze Loss, reporting superior segmentation across ten organs on multiple modalities.
-
Feature-Aware Anisotropic Local Differential Privacy for Utility-Preserving Graph Representation Learning in Metal Additive Manufacturing
FI-LDP-HGAT applies feature-importance-aware anisotropic local differential privacy to a hierarchical graph attention network, recovering 81.5% utility at epsilon=4 and 0.762 defect recall at epsilon=2 on a DED porosity dataset while outperforming standard LDP and DP-SGD baselines.
-
OASIC: Occlusion-Agnostic and Severity-Informed Classification
OASIC uses anomaly-based masking and severity estimation to select occlusion-matched models, improving AUC on occluded images by up to 23.7 points.
-
Can LLMs Learn to Reason Robustly under Noisy Supervision?
Online Label Refinement lets LLMs learn robust reasoning from noisy supervision by correcting labels when majority answers show rising rollout success and stable history, delivering 3-4% gains on math and reasoning benchmarks even at high noise levels.
-
R\'enyi Attention Entropy for Patch Pruning
Rényi entropy of attention maps serves as a tunable criterion for pruning redundant patches in vision transformers, reducing compute with preserved accuracy on image recognition.
-
Mitigating Shortcut Learning via Feature Disentanglement in Medical Imaging: A Benchmark Study
Benchmark shows that combining data rebalancing with feature disentanglement mitigates shortcut learning more effectively than rebalancing alone in medical imaging models.
-
Improving Backward Conformal Prediction via Non-Conformity Score Transformation
A step-function score transformation I_w = w·1{s≥w} reduces the estimated-coverage gap in Backward Conformal Prediction from about 4.2% to 1.1% on CIFAR-10, CIFAR-100, and Tiny-ImageNet.
-
A Retrieval-Augmented Generation Approach to Extracting Algorithmic Logic from Neural Networks
NN-RAG extracts 1,289 candidate neural modules from 19 PyTorch repositories, validates 941 of them, and supplies roughly 72% of the novel structures in the LEMUR dataset while enabling cross-repository migration.
-
LoFT: Parameter-Efficient Fine-Tuning for Long-tailed Semi-Supervised Learning in Open-World Scenarios
LoFT uses parameter-efficient fine-tuning of foundation models for long-tailed semi-supervised learning, supported by proofs that this reduces hypothesis complexity to minimize balanced posterior error and compresses outlier acceptance regions, with LoFT-OW handling open-world OOD cases.
-
Normalized Matching Transformer
Normalized Matching Transformer enforces unit-norm embeddings at every Transformer layer and trains with InfoNCE plus hyperspherical uniformity loss, reaching new state-of-the-art accuracy on PascalVOC and SPair-71k while converging faster than prior matching networks.
-
YOLOv12: Attention-Centric Real-Time Object Detectors
YOLOv12 is a new attention-based real-time object detector that reports higher accuracy than YOLOv10, YOLOv11, and RT-DETR variants at comparable or better speed and efficiency.
-
SpikeDet: Better Firing Patterns for Accurate and Energy-Efficient Object Detection with Spiking Neural Networks
SpikeDet reaches 52.2% AP on COCO 2017 with spiking networks by optimizing firing patterns via MDSNet and SMFM, using half the energy of prior SNN detectors.
-
Revisiting Feature Prediction for Learning Visual Representations from Video
V-JEPA models trained only on feature prediction from 2 million public videos achieve 81.9% on Kinetics-400, 72.2% on Something-Something-v2, and 77.9% on ImageNet-1K using frozen ViT-H/16 backbones.
-
Sharpness-Aware Minimization for Efficiently Improving Generalization
SAM solves a min-max problem to locate flat low-loss regions, improving generalization on CIFAR, ImageNet and label-noise tasks.
-
Learning Data Augmentation Strategies for Object Detection
Learned data augmentation policies optimized for object detection improve COCO mAP by more than 2.3 and transfer to other datasets and models.
-
Benchmarking the Robustness of Autonomous Driving to Environmental Illusions: A Lane Perception Perspective
Environmental illusions cause 5-7% accuracy drops in lane detection models and can trigger collisions in closed-loop simulation, with a proposed defense (MIDA) recovering ~4% robustness.
-
MedDiffuseMix: Preserving Diagnostic Evidence with Saliency-Aware Diffusion Medical Image Data Augmentation
Saliency-guided diffusion mixing that preserves Grad-CAM-highlighted diagnostic regions improves medical image classification accuracy and AUC across four public datasets.
-
Distill Once, Adapt Life-Long: Exploring Dataset Distillation for Continual Test-Time Adaptation
DO-ALL applies dataset distillation to generate synthetic source anchors that stabilize continual test-time adaptation under evolving domains without storing original source data.