mixup: Beyond Empirical Risk Minimization

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.

ICLR camera ready version. Changes vs V1: fix repo URL; add ablation studies; add mixup + dropout etc

References in corpus (5)

• SMOTE: Synthetic Minority Over-sampling Technique
• Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
• A Closer Look at Memorization in Deep Networks
• Spectrally-normalized margin bounds for neural networks
• Dataset Augmentation in Feature Space

Cited by in corpus (787)

• YOLOv4: Optimal Speed and Accuracy of Object Detection
• PVT v2: Improved Baselines with Pyramid Vision Transformer
• FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
• Neural Architecture Search: A Survey
• AutoML: A Survey of the State-of-the-Art
• Unsupervised Data Augmentation for Consistency Training
• Recent advances and clinical applications of deep learning in medical image analysis
• Transformer in Transformer
• fastai: A Layered API for Deep Learning
• Deep Neural Networks and Tabular Data: A Survey
• CoAtNet: Marrying Convolution and Attention for All Data Sizes
• AutoAugment: Learning Augmentation Policies from Data
• PadChest: A large chest x-ray image dataset with multi-label annotated reports
• A Comprehensive Survey of Image Augmentation Techniques for Deep Learning
• CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
• MixMatch: A Holistic Approach to Semi-Supervised Learning
• AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
• Efficient Deep Learning: A Survey on Making Deep Learning Models Smaller, Faster, and Better
• ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision
• Data Augmentation using Random Image Cropping and Patching for Deep CNNs
• Learning Spatial Fusion for Single-Shot Object Detection
• A Survey on Data Augmentation for Text Classification
• Avoiding Overfitting: A Survey on Regularization Methods for Convolutional Neural Networks
• Data Augmentation Approaches in Natural Language Processing: A Survey
• Image Classification with Deep Learning in the Presence of Noisy Labels: A Survey
• DeepViT: Towards Deeper Vision Transformer
• A Survey of Sound Source Localization with Deep Learning Methods
• VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text
• Data Augmentation by Pairing Samples for Images Classification
• Billion-scale semi-supervised learning for image classification
• Single-source Domain Expansion Network for Cross-Scene Hyperspectral Image Classification
• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• A Simple Semi-Supervised Learning Framework for Object Detection
• Escaping the Big Data Paradigm with Compact Transformers
• Image Quality Assessment using Contrastive Learning
• Attention Bottlenecks for Multimodal Fusion
• Out-of-Distribution Generalization via Risk Extrapolation (REx)
• High-Performance Large-Scale Image Recognition Without Normalization
• An Overview of Deep Semi-Supervised Learning
• GridMask Data Augmentation
• An Empirical Study of Remote Sensing Pretraining
• Efficient Training of Audio Transformers with Patchout
• A survey on Semi-, Self- and Unsupervised Learning for Image Classification
• A feature agnostic approach for glaucoma detection in OCT volumes
• Revisiting ResNets: Improved Training and Scaling Strategies
• DoFE: Domain-oriented Feature Embedding for Generalizable Fundus Image Segmentation on Unseen Datasets
• A Comprehensive Survey of Neural Architecture Search: Challenges and Solutions
• Target-Speaker Voice Activity Detection: a Novel Approach for Multi-Speaker Diarization in a Dinner Party Scenario
• You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection
• BadNL: Backdoor Attacks against NLP Models with Semantic-preserving Improvements
• ShakeDrop Regularization for Deep Residual Learning
• A Survey of Zero-shot Generalisation in Deep Reinforcement Learning
• NATS-Bench: Benchmarking NAS Algorithms for Architecture Topology and Size
• ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
• Automatic Brain Tumor Segmentation using Convolutional Neural Networks with Test-Time Augmentation
• Automated data processing and feature engineering for deep learning and big data applications: a survey
• Data Augmentation in Natural Language Processing: A Novel Text Generation Approach for Long and Short Text Classifiers
• MicroNets: Neural Network Architectures for Deploying TinyML Applications on Commodity Microcontrollers
• Bag of Freebies for Training Object Detection Neural Networks
• Augmenting Data with Mixup for Sentence Classification: An Empirical Study
• All Tokens Matter: Token Labeling for Training Better Vision Transformers
• Bag of Tricks for Image Classification with Convolutional Neural Networks
• Overview and Evaluation of Sound Event Localization and Detection in DCASE 2019
• High-throughput molecular imaging via deep learning enabled Raman spectroscopy
• A New Dataset, Poisson GAN and AquaNet for Underwater Object Grabbing
• HRFormer: High-Resolution Transformer for Dense Prediction
• Recent Progress in Transformer-based Medical Image Analysis
• TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-captured Scenarios
• SAINT: Improved Neural Networks for Tabular Data via Row Attention and Contrastive Pre-Training
• Improve Unsupervised Domain Adaptation with Mixup Training
• Heterogeneous Domain Generalization via Domain Mixup
• DPT: Deformable Patch-based Transformer for Visual Recognition
• Supervised Contrastive Learning
• Privacy-Preserving Deep Learning via Weight Transmission
• Fixup Initialization: Residual Learning Without Normalization
• PP-OCR: A Practical Ultra Lightweight OCR System
• Adversarial Attacks Against Medical Deep Learning Systems
• LongReMix: Robust Learning with High Confidence Samples in a Noisy Label Environment
• Memory-aware curriculum federated learning for breast cancer classification
• ByteTrack: Multi-Object Tracking by Associating Every Detection Box
• Mixing Up Contrastive Learning: Self-Supervised Representation Learning for Time Series
• FMix: Enhancing Mixed Sample Data Augmentation
• Sharpness-Aware Minimization for Efficiently Improving Generalization
• Integrating Expert Knowledge with Domain Adaptation for Unsupervised Fault Diagnosis
• CMID: A Unified Self-Supervised Learning Framework for Remote Sensing Image Understanding
• CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows
• CycleMLP: A MLP-like Architecture for Dense Prediction
• Universal Adversarial Examples in Remote Sensing: Methodology and Benchmark
• Data augmentation instead of explicit regularization
• Review of data analysis in vision inspection of power lines with an in-depth discussion of deep learning technology
• PP-YOLOv2: A Practical Object Detector
• Improving Robustness Without Sacrificing Accuracy with Patch Gaussian Augmentation
• AADG: Automatic Augmentation for Domain Generalization on Retinal Image Segmentation
• SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularization
• Long-tailed Visual Recognition via Gaussian Clouded Logit Adjustment
• Self-Adaptive Training: beyond Empirical Risk Minimization
• Towards a universal neural network encoder for time series
• In Search of Lost Domain Generalization
• An overview of mixing augmentation methods and augmentation strategies
• Domain Generalization for Medical Imaging Classification with Linear-Dependency Regularization
• BBN: Bilateral-Branch Network with Cumulative Learning for Long-Tailed Visual Recognition
• DSP: Dual Soft-Paste for Unsupervised Domain Adaptive Semantic Segmentation
• PARSE: Pairwise Alignment of Representations in Semi-Supervised EEG Learning for Emotion Recognition
• Reference-guided Pseudo-Label Generation for Medical Semantic Segmentation
• Training Binary Neural Networks with Real-to-Binary Convolutions
• Knowledge Distillation in Generations: More Tolerant Teachers Educate Better Students
• Towards Lightweight Transformer via Group-wise Transformation for Vision-and-Language Tasks
• Efficient Multi-objective Neural Architecture Search via Lamarckian Evolution
• RepMLP: Re-parameterizing Convolutions into Fully-connected Layers for Image Recognition
• Patch-Mix Contrastive Learning with Audio Spectrogram Transformer on Respiratory Sound Classification
• A Systematic Survey of Regularization and Normalization in GANs
• ConTNet: Why not use convolution and transformer at the same time?
• Increasing Diversity While Maintaining Accuracy: Text Data Generation with Large Language Models and Human Interventions
• Peer Loss Functions: Learning from Noisy Labels without Knowing Noise Rates
• SOFT: Softmax-free Transformer with Linear Complexity
• Convolutional Neural Networks with Gated Recurrent Connections
• BYOL for Audio: Exploring Pre-trained General-purpose Audio Representations
• MRI-based Alzheimer's disease prediction via distilling the knowledge in multi-modal data
• Towards Utilizing Unlabeled Data in Federated Learning: A Survey and Prospective
• A Survey of Deep Learning for Scientific Discovery
• Remix: Rebalanced Mixup
• Receptive Field Regularization Techniques for Audio Classification and Tagging with Deep Convolutional Neural Networks
• Noise Contrastive Priors for Functional Uncertainty
• Vision Transformer Pruning
• MixCo: Mix-up Contrastive Learning for Visual Representation
• Affinity and Diversity: Quantifying Mechanisms of Data Augmentation
• Understanding and Improving Early Stopping for Learning with Noisy Labels
• Self-Distillation as Instance-Specific Label Smoothing
• Multi-Sample Dropout for Accelerated Training and Better Generalization
• Understanding the Failure Modes of Out-of-Distribution Generalization
• Depth-Wise Convolutions in Vision Transformers for Efficient Training on Small Datasets
• RobustART: Benchmarking Robustness on Architecture Design and Training Techniques
• Augment your batch: better training with larger batches
• Convolutional Fine-Grained Classification with Self-Supervised Target Relation Regularization
• The DeepFake Detection Challenge (DFDC) Dataset
• Embrace Limited and Imperfect Training Datasets: Opportunities and Challenges in Plant Disease Recognition Using Deep Learning
• MEAL V2: Boosting Vanilla ResNet-50 to 80%+ Top-1 Accuracy on ImageNet without Tricks
• Feature Fusion from Head to Tail for Long-Tailed Visual Recognition
• VideoMix: Rethinking Data Augmentation for Video Classification
• On the Effects of Different Types of Label Noise in Multi-Label Remote Sensing Image Classification
• Self-Augmentation: Generalizing Deep Networks to Unseen Classes for Few-Shot Learning
• Structured Consistency Loss for semi-supervised semantic segmentation
• A Two-Stage Approach to Device-Robust Acoustic Scene Classification
• Device-Robust Acoustic Scene Classification Based on Two-Stage Categorization and Data Augmentation
• RepVGG: Making VGG-style ConvNets Great Again
• Improving Generalization in Reinforcement Learning with Mixture Regularization
• Uncertainty-Induced Transferability Representation for Source-Free Unsupervised Domain Adaptation
• Conservative-Progressive Collaborative Learning for Semi-supervised Semantic Segmentation
• Model Patching: Closing the Subgroup Performance Gap with Data Augmentation
• Overfitting in adversarially robust deep learning
• KD3A: Unsupervised Multi-Source Decentralized Domain Adaptation via Knowledge Distillation
• Deep Weakly-Supervised Learning Methods for Classification and Localization in Histology Images: A Survey
• Towards long-tailed, multi-label disease classification from chest X-ray: Overview of the CXR-LT challenge
• High Frequency Component Helps Explain the Generalization of Convolutional Neural Networks
• Recent Advances in Adversarial Training for Adversarial Robustness
• The LHC Olympics 2020: A Community Challenge for Anomaly Detection in High Energy Physics
• Greedy Policy Search: A Simple Baseline for Learnable Test-Time Augmentation
• Vision Transformers with Patch Diversification
• Learning perturbation sets for robust machine learning
• Learning Invariance from Generated Variance for Unsupervised Person Re-identification
• Underwater Acoustic Target Recognition based on Smoothness-inducing Regularization and Spectrogram-based Data Augmentation
• Acoustic Scene Classification with Squeeze-Excitation Residual Networks
• MS-DETR: Multispectral Pedestrian Detection Transformer with Loosely Coupled Fusion and Modality-Balanced Optimization
• Learning Loss for Test-Time Augmentation
• FINE Samples for Learning with Noisy Labels
• UniformAugment: A Search-free Probabilistic Data Augmentation Approach
• Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks
• Ensemble learning in CNN augmented with fully connected subnetworks
• Continual Semi-Supervised Learning through Contrastive Interpolation Consistency
• SemiFL: Semi-Supervised Federated Learning for Unlabeled Clients with Alternate Training
• Class-Imbalanced Semi-Supervised Learning
• An Overview of Indian Spoken Language Recognition from Machine Learning Perspective
• Cut-Thumbnail: A Novel Data Augmentation for Convolutional Neural Network
• Contrastive Learning with Stronger Augmentations
• An Empirical Evaluation on Robustness and Uncertainty of Regularization Methods
• Rethinking Distributional Matching Based Domain Adaptation
• Compounding the Performance Improvements of Assembled Techniques in a Convolutional Neural Network
• HS-ResNet: Hierarchical-Split Block on Convolutional Neural Network
• Confidence Scores Make Instance-dependent Label-noise Learning Possible
• General audio tagging with ensembling convolutional neural network and statistical features
• A Comprehensive and Modularized Statistical Framework for Gradient Norm Equality in Deep Neural Networks
• PatchUp: A Feature-Space Block-Level Regularization Technique for Convolutional Neural Networks
• PSLT: A Light-weight Vision Transformer with Ladder Self-Attention and Progressive Shift
• REB: Reducing Biases in Representation for Industrial Anomaly Detection
• ResNet strikes back: An improved training procedure in timm
• Glance-and-Gaze Vision Transformer
• L_DMI: An Information-theoretic Noise-robust Loss Function
• Towards Better Accuracy-efficiency Trade-offs: Divide and Co-training
• Meta-FDMixup: Cross-Domain Few-Shot Learning Guided by Labeled Target Data
• Fully Transformer Networks for Semantic Image Segmentation
• PConv: Simple yet Effective Convolutional Layer for Generative Adversarial Network
• LEAF: A Learnable Frontend for Audio Classification
• Learning to Defend by Learning to Attack
• Improving Auto-Augment via Augmentation-Wise Weight Sharing
• CoDA: Contrast-enhanced and Diversity-promoting Data Augmentation for Natural Language Understanding
• QMaxViT-Unet+: A Query-Based MaxViT-Unet with Edge Enhancement for Scribble-Supervised Segmentation of Medical Images
• Large-Scale Generative Data-Free Distillation
• Improving COVID-19 CT Classification of CNNs by Learning Parameter-Efficient Representation
• Multi-Granularity Regularized Re-Balancing for Class Incremental Learning
• Data Augmentation for Meta-Learning
• Assessing Post-Disaster Damage from Satellite Imagery using Semi-Supervised Learning Techniques
• Sequence-to-Sequence Data Augmentation for Dialogue Language Understanding
• Interpolated Adversarial Training: Achieving Robust Neural Networks without Sacrificing Too Much Accuracy
• Deep Open-Set Recognition for Silicon Wafer Production Monitoring
• Automated Human Cell Classification in Sparse Datasets using Few-Shot Learning
• Uncertainty-Aware Reliable Text Classification
• Semi-Supervised Building Footprint Generation with Feature and Output Consistency Training
• Semi-Federated Learning: Convergence Analysis and Optimization of A Hybrid Learning Framework
• Dynamic Scale Training for Object Detection
• ConvMLP: Hierarchical Convolutional MLPs for Vision
• CSL-YOLO: A New Lightweight Object Detection System for Edge Computing
• Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition
• MixupExplainer: Generalizing Explanations for Graph Neural Networks with Data Augmentation
• A Four-Stage Data Augmentation Approach to ResNet-Conformer Based Acoustic Modeling for Sound Event Localization and Detection
• TGDM: Target Guided Dynamic Mixup for Cross-Domain Few-Shot Learning
• Rethinking and Improving Relative Position Encoding for Vision Transformer
• Domain Disentanglement with Interpolative Data Augmentation for Dual-Target Cross-Domain Recommendation
• AclNet: efficient end-to-end audio classification CNN
• Diverse Branch Block: Building a Convolution as an Inception-like Unit
• Sample Mixed-Based Data Augmentation for Domestic Audio Tagging
• Toward Degree Bias in Embedding-Based Knowledge Graph Completion
• VOLO: Vision Outlooker for Visual Recognition
• CarveMix: A Simple Data Augmentation Method for Brain Lesion Segmentation
• Graph Random Neural Network for Semi-Supervised Learning on Graphs
• Fixing the train-test resolution discrepancy
• A Baseline for Few-Shot Image Classification
• Meta Transition Adaptation for Robust Deep Learning with Noisy Labels
• Inverse Image Frequency for Long-tailed Image Recognition
• Anatomy of Catastrophic Forgetting: Hidden Representations and Task Semantics
• RAIN: RegulArization on Input and Network for Black-Box Domain Adaptation
• Sill-Net: Feature Augmentation with Separated Illumination Representation
• Viewmaker Networks: Learning Views for Unsupervised Representation Learning
• Uncertainty-Guided Mixup for Semi-Supervised Domain Adaptation without Source Data
• Deep Learning for Classification and Severity Estimation of Coffee Leaf Biotic Stress
• Reinforcing Generated Images via Meta-learning for One-Shot Fine-Grained Visual Recognition
• An Empirical Survey of Data Augmentation for Limited Data Learning in NLP
• Feature Combination Meets Attention: Baidu Soccer Embeddings and Transformer based Temporal Detection
• Selecting Data Augmentation for Simulating Interventions
• ConvBoost: Boosting ConvNets for Sensor-based Activity Recognition
• EViT: An Eagle Vision Transformer with Bi-Fovea Self-Attention
• Backdoor Attacks for Remote Sensing Data with Wavelet Transform
• GuidedMixup: An Efficient Mixup Strategy Guided by Saliency Maps
• MVA2023 Small Object Detection Challenge for Spotting Birds: Dataset, Methods, and Results
• Multi-Grained Angle Representation for Remote Sensing Object Detection
• A Unified Survey on Anomaly, Novelty, Open-Set, and Out-of-Distribution Detection: Solutions and Future Challenges
• Deep Learning to Estimate Permeability using Geophysical Data
• AIM 2020 Challenge on Real Image Super-Resolution: Methods and Results
• Fair Oversampling Technique using Heterogeneous Clusters
• Interpreting Adversarial Robustness: A View from Decision Surface in Input Space
• Semi-Supervised Sequence Modeling with Cross-View Training
• A dual-branch model with inter- and intra-branch contrastive loss for long-tailed recognition
• What Matters for Adversarial Imitation Learning?
• Robust Duality Learning for Unsupervised Visible-Infrared Person Re-Identification
• Combining MixMatch and Active Learning for Better Accuracy with Fewer Labels
• Vicinal and categorical domain adaptation
• Omni-sourced Webly-supervised Learning for Video Recognition
• PointAugment: an Auto-Augmentation Framework for Point Cloud Classification
• Rainbow Memory: Continual Learning with a Memory of Diverse Samples
• Few Shot Network Compression via Cross Distillation
• Contrastive Learning for Compact Single Image Dehazing
• PseudoAugment: Learning to Use Unlabeled Data for Data Augmentation in Point Clouds
• DP-InstaHide: Provably Defusing Poisoning and Backdoor Attacks with Differentially Private Data Augmentations
• MetaSAug: Meta Semantic Augmentation for Long-Tailed Visual Recognition
• On the effectiveness of adversarial training against common corruptions
• RETRIEVE: Coreset Selection for Efficient and Robust Semi-Supervised Learning
• HardCoRe-NAS: Hard Constrained diffeRentiable Neural Architecture Search
• Defending Against Adversarial Examples with K-Nearest Neighbor
• Dirichlet-Based Prediction Calibration for Learning with Noisy Labels
• ELF: An Early-Exiting Framework for Long-Tailed Classification
• Refine Myself by Teaching Myself: Feature Refinement via Self-Knowledge Distillation
• ConvNets and ImageNet Beyond Accuracy: Understanding Mistakes and Uncovering Biases
• LaplaceNet: A Hybrid Graph-Energy Neural Network for Deep Semi-Supervised Classification
• Artificial and beneficial -- Exploiting artificial images for aerial vehicle detection
• AutoFormer: Searching Transformers for Visual Recognition
• Flow-Mixup: Classifying Multi-labeled Medical Images with Corrupted Labels
• On Data-Augmentation and Consistency-Based Semi-Supervised Learning
• Self-supervised Pre-training with Hard Examples Improves Visual Representations
• Towards Strengthening Deep Learning-based Side Channel Attacks with Mixup
• RoIMix: Proposal-Fusion among Multiple Images for Underwater Object Detection
• Improving Semi-supervised Federated Learning by Reducing the Gradient Diversity of Models
• Representation Based Complexity Measures for Predicting Generalization in Deep Learning
• MixUp as Directional Adversarial Training
• FoPro-KD: Fourier Prompted Effective Knowledge Distillation for Long-Tailed Medical Image Recognition
• Accurate identification and measurement of the precipitate area by two-stage deep neural networks in novel chromium-based alloys
• A Large Multi-Target Dataset of Common Bengali Handwritten Graphemes
• Joint framework with deep feature distillation and adaptive focal loss for weakly supervised audio tagging and acoustic event detection
• SimAug: Learning Robust Representations from Simulation for Trajectory Prediction
• Semi-supervised Models are Strong Unsupervised Domain Adaptation Learners
• MixStyle Neural Networks for Domain Generalization and Adaptation
• SimMIM: A Simple Framework for Masked Image Modeling
• Universum-inspired Supervised Contrastive Learning
• Multimodal Urban Sound Tagging with Spatiotemporal Context
• Efficient Differentiable Neural Architecture Search with Meta Kernels
• Music theme recognition using CNN and self-attention
• Analyze Mass Spectrometry data with Artificial Intelligence to assist the understanding of past habitability of Mars and provide insights for future missions
• Physics-separating artificial neural networks for predicting initial stages of Al sputtering and thin film deposition in Ar plasma discharges
• A New Semi-supervised Learning Benchmark for Classifying View and Diagnosing Aortic Stenosis from Echocardiograms
• Learning Task-relevant Representations for Generalization via Characteristic Functions of Reward Sequence Distributions
• Cross-Referencing Self-Training Network for Sound Event Detection in Audio Mixtures
• Dual Mixup Regularized Learning for Adversarial Domain Adaptation
• Improving Perceptual Quality of Drum Transcription with the Expanded Groove MIDI Dataset
• Boosting Discriminative Visual Representation Learning with Scenario-Agnostic Mixup
• Understanding Open-Set Recognition by Jacobian Norm and Inter-Class Separation
• SECANT: Self-Expert Cloning for Zero-Shot Generalization of Visual Policies
• An Empirical Analysis of the Impact of Data Augmentation on Knowledge Distillation
• Mine Your Own vieW: Self-Supervised Learning Through Across-Sample Prediction
• Dimension Independent Mixup for Hard Negative Sample in Collaborative Filtering
• Maximizing Discrimination Capability of Knowledge Distillation with Energy Function
• GraphMixup: Improving Class-Imbalanced Node Classification on Graphs by Self-supervised Context Prediction
• Multi-organ segmentation: a progressive exploration of learning paradigms under scarce annotation
• Gradual Domain Adaptation without Indexed Intermediate Domains
• Gated Convolutional Networks with Hybrid Connectivity for Image Classification
• Co-Learning Meets Stitch-Up for Noisy Multi-label Visual Recognition
• Robust Image Classification Using A Low-Pass Activation Function and DCT Augmentation
• Learning data augmentation policies using augmented random search
• Paraphrase Augmented Task-Oriented Dialog Generation
• A Competitive Method to VIPriors Object Detection Challenge
• Diverse Ensembles Improve Calibration
• ConDA: Continual Unsupervised Domain Adaptation
• Domain Adaptive Detection of MAVs: A Benchmark and Noise Suppression Network
• On the Generalization Effects of Linear Transformations in Data Augmentation
• Few-shot Forgery Detection via Guided Adversarial Interpolation
• Spectrum Correction: Acoustic Scene Classification with Mismatched Recording Devices
• Diabetic Foot Ulcer Grand Challenge 2022 Summary
• Incorporating Anatomical Awareness for Enhanced Generalizability and Progression Prediction in Deep Learning-Based Radiographic Sacroiliitis Detection
• Distance-Based Learning from Errors for Confidence Calibration
• Simulations of Common Unsupervised Domain Adaptation Algorithms for Image Classification
• Robust Mutual Learning for Semi-supervised Semantic Segmentation
• Building Surrogate Models of Nuclear Density Functional Theory with Gaussian Processesand Autoencoders
• SpecMix : A Mixed Sample Data Augmentation method for Training withTime-Frequency Domain Features
• bLIMEy: Surrogate Prediction Explanations Beyond LIME
• Dens-PU: PU Learning with Density-Based Positive Labeled Augmentation
• Better Modelling Out-of-Distribution Regression on Distributed Acoustic Sensor Data Using Anchored Hidden State Mixup
• Improving the Performance of Automated Audio Captioning via Integrating the Acoustic and Semantic Information
• Dataset balancing can hurt model performance
• Acoustic Scene Classification with Spectrogram Processing Strategies
• Webly Supervised Image Classification with Metadata: Automatic Noisy Label Correction via Visual-Semantic Graph
• Is Private Learning Possible with Instance Encoding?
• Improving Medical Image Classification with Label Noise Using Dual-uncertainty Estimation
• VIPriors 1: Visual Inductive Priors for Data-Efficient Deep Learning Challenges
• ACE: Ally Complementary Experts for Solving Long-Tailed Recognition in One-Shot
• Learning outside the Black-Box: The pursuit of interpretable models
• GP: Context-free Grammar Pre-training for Text-to-SQL Parsers
• Spoken Language Identification using ConvNets
• Comparing to Learn: Surpassing ImageNet Pretraining on Radiographs By Comparing Image Representations
• MixMatch Domain Adaptaion: Prize-winning solution for both tracks of VisDA 2019 challenge
• Proximity Sensing: Modeling and Understanding Noisy RSSI-BLE Signals and Other Mobile Sensor Data for Digital Contact Tracing
• Self-Paced Contrastive Learning for Semi-supervised Medical Image Segmentation with Meta-labels
• Understand and Improve Contrastive Learning Methods for Visual Representation: A Review
• SCARF: Self-Supervised Contrastive Learning using Random Feature Corruption
• Contrastive Learning Improves Model Robustness Under Label Noise
• Unsupervised Anomaly Detection with Adversarial Mirrored AutoEncoders
• Retrieval-Augmented Convolutional Neural Networks for Improved Robustness against Adversarial Examples
• FenceMask: A Data Augmentation Approach for Pre-extracted Image Features
• Image Obfuscation for Privacy-Preserving Machine Learning
• Towards recognizing the light facet of the Higgs Boson
• Distributed Semi-supervised Fuzzy Regression with Interpolation Consistency Regularization
• Guided Variational Autoencoder for Disentanglement Learning
• Cyclic Differentiable Architecture Search
• Self-training with noisy student model and semi-supervised loss function for dcase 2021 challenge task 4
• BitMix: Data Augmentation for Image Steganalysis
• ZerO Initialization: Initializing Neural Networks with only Zeros and Ones
• Identifying the Defective: Detecting Damaged Grains for Cereal Appearance Inspection
• Attention based Convolutional Recurrent Neural Network for Environmental Sound Classification
• Detecting Adversarial Examples in Convolutional Neural Networks
• Robust Semi-Supervised Learning for Self-learning Open-World Classes
• Generalizable Pancreas Segmentation via a Dual Self-Supervised Learning Framework
• Contrastive-mixup learning for improved speaker verification
• Iterative Graph Self-Distillation
• Deep learning in magnetic resonance prostate segmentation: A review and a new perspective
• Physics-Informed Deep Learning and Partial Transfer Learning for Bearing Fault Diagnosis in the Presence of Highly Missing Data
• In Search of netUnicorn: A Data-Collection Platform to Develop Generalizable ML Models for Network Security Problems
• Online Hyper-parameter Learning for Auto-Augmentation Strategy
• Sleeper Agent: Scalable Hidden Trigger Backdoors for Neural Networks Trained from Scratch
• A locality-based approach for coded computation
• A Bird Song Detector for improving bird identification through Deep Learning: a case study from Doñana
• Improving Calibration for Long-Tailed Recognition
• Autoencoder Image Interpolation by Shaping the Latent Space
• Hijack Vertical Federated Learning Models As One Party
• Augmentation Strategies for Learning with Noisy Labels
• Leave Zero Out: Towards a No-Cross-Validation Approach for Model Selection
• Should Ensemble Members Be Calibrated?
• Rethinking Token-Mixing MLP for MLP-based Vision Backbone
• Fine-grained Synthesis of Unrestricted Adversarial Examples
• Triple Generative Adversarial Networks
• Semi-Supervised Domain Generalizable Person Re-Identification
• Towards Robust Vision Transformer
• Semi-supervised Domain Adaptation based on Dual-level Domain Mixing for Semantic Segmentation
• Class-Distribution-Aware Calibration for Long-Tailed Visual Recognition
• Deep Learning in Diabetic Foot Ulcers Detection: A Comprehensive Evaluation
• It Takes Two to Tango: Mixup for Deep Metric Learning
• Improving Unsupervised Image Clustering With Robust Learning
• Adversarial Feature Augmentation and Normalization for Visual Recognition
• ReMix: Calibrated Resampling for Class Imbalance in Deep learning
• Geometric Understanding of Discriminability and Transferability for Visual Domain Adaptation
• Context Decoupling Augmentation for Weakly Supervised Semantic Segmentation
• Coresets for Robust Training of Neural Networks against Noisy Labels
• Data Interpolating Prediction: Alternative Interpretation of Mixup
• KuroNet: Pre-Modern Japanese Kuzushiji Character Recognition with Deep Learning
• From Synthetic to Real: Unsupervised Domain Adaptation for Animal Pose Estimation
• XLST: Cross-lingual Self-training to Learn Multilingual Representation for Low Resource Speech Recognition
• Evaluating COPY-BLEND Augmentation for Low Level Vision Tasks
• Cross-directional Feature Fusion Network for Building Damage Assessment from Satellite Imagery
• MSE Loss with Outlying Label for Imbalanced Classification
• Federated Learning on Non-IID Data: A Survey
• Towards Training Stronger Video Vision Transformers for EPIC-KITCHENS-100 Action Recognition
• Adversarial Visual Robustness by Causal Intervention
• Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training
• Uni-Perceiver: Pre-training Unified Architecture for Generic Perception for Zero-shot and Few-shot Tasks
• MMFL-Net: Multi-scale and Multi-granularity Feature Learning for Cross-domain Fashion Retrieval
• Cycle Self-Training for Semi-Supervised Object Detection with Distribution Consistency Reweighting
• Parameterization of Cross-Token Relations with Relative Positional Encoding for Vision MLP
• Leveraging Angular Distributions for Improved Knowledge Distillation
• SST: Self-training with Self-adaptive Thresholding for Semi-supervised Learning
• YOLO11-JDE: Fast and Accurate Multi-Object Tracking with Self-Supervised Re-ID
• DPGIIL: Dirichlet Process-Deep Generative Model-Integrated Incremental Learning for Clustering in Transmissibility-based Online Structural Anomaly Detection
• Supervision by Denoising for Medical Image Segmentation
• FakeMix Augmentation Improves Transparent Object Detection
• Supporting Clustering with Contrastive Learning
• When and How Mixup Improves Calibration
• Identifying Mislabeled Images in Supervised Learning Utilizing Autoencoder
• Adaptive Label Smoothing
• Learning to Impute: A General Framework for Semi-supervised Learning
• Prune and Replace NAS
• A Simple Cache Model for Image Recognition
• Deep Semi-Supervised Learning for Time Series Classification
• A3T: Adversarially Augmented Adversarial Training
• CNNs Avoid Curse of Dimensionality by Learning on Patches
• What Can be Seen is What You Get: Structure Aware Point Cloud Augmentation
• Semi-supervised Viewpoint Estimation with Geometry-aware Conditional Generation
• Adversarial Attacks in Sound Event Classification
• Laplacian Networks: Bounding Indicator Function Smoothness for Neural Network Robustness
• LaSO: Label-Set Operations networks for multi-label few-shot learning
• Confidence Adaptive Regularization for Deep Learning with Noisy Labels
• Embedding Expansion: Augmentation in Embedding Space for Deep Metric Learning
• When Human Pose Estimation Meets Robustness: Adversarial Algorithms and Benchmarks
• Reliable Student: Addressing Noise in Semi-Supervised 3D Object Detection
• Combining Self-Supervised and Supervised Learning with Noisy Labels
• Probing the Purview of Neural Networks via Gradient Analysis
• OpenCoS: Contrastive Semi-supervised Learning for Handling Open-set Unlabeled Data
• Revisiting Image Aesthetic Assessment via Self-Supervised Feature Learning
• Between-class Learning for Image Classification
• Weakly Supervised Disentangled Representation for Goal-conditioned Reinforcement Learning
• CAPE: Encoding Relative Positions with Continuous Augmented Positional Embeddings
• On The State of Data In Computer Vision: Human Annotations Remain Indispensable for Developing Deep Learning Models
• Parameter Exchange for Robust Dynamic Domain Generalization
• On Data Augmentation and Adversarial Risk: An Empirical Analysis
• Calibrate and Prune: Improving Reliability of Lottery Tickets Through Prediction Calibration
• Meta Approach to Data Augmentation Optimization
• Augmentation Pathways Network for Visual Recognition
• Towards Robustness to Label Noise in Text Classification via Noise Modeling
• PAFNet: An Efficient Anchor-Free Object Detector Guidance
• MixSpeech: Data Augmentation for Low-resource Automatic Speech Recognition
• Revisiting Knowledge Distillation for Object Detection
• A Fusion-Denoising Attack on InstaHide with Data Augmentation
• Model-Based Domain Generalization
• Adaptive Sample Selection for Robust Learning under Label Noise
• Computation-Efficient Knowledge Distillation via Uncertainty-Aware Mixup
• A Loss Curvature Perspective on Training Instability in Deep Learning
• Generalization of Reinforcement Learning with Policy-Aware Adversarial Data Augmentation
• Catch-Up Mix: Catch-Up Class for Struggling Filters in CNN
• Parallel Grid Pooling for Data Augmentation
• Vision-Based Uncertainty-Aware Motion Planning based on Probabilistic Semantic Segmentation
• E-Stitchup: Data Augmentation for Pre-Trained Embeddings
• Physics-separating artificial neural networks for predicting sputtering and thin film deposition of AlN in Ar/N${}_2$ discharges on experimental timescales
• Combining High-Level Features of Raw Audio Waves and Mel-Spectrograms for Audio Tagging
• Practical and Secure Federated Recommendation with Personalized Masks
• Jigsaw-VAE: Towards Balancing Features in Variational Autoencoders
• Hire-MLP: Vision MLP via Hierarchical Rearrangement
• Generalization Bounds for Vicinal Risk Minimization Principle
• Spatial mixup: Directional loudness modification as data augmentation for sound event localization and detection
• More Than Meets The Eye: Semi-supervised Learning Under Non-IID Data
• Negative sampling in semi-supervised learning
• Deep Convolutional Neural Network with Mixup for Environmental Sound Classification
• Neural Complexity Measures
• Quantile Regularization: Towards Implicit Calibration of Regression Models
• Machine learning with limited data
• ObjectAug: Object-level Data Augmentation for Semantic Image Segmentation
• Robust and On-the-fly Dataset Denoising for Image Classification
• Drop-Activation: Implicit Parameter Reduction and Harmonic Regularization
• Recent Deep Semi-supervised Learning Approaches and Related Works
• ACGAN-based Data Augmentation Integrated with Long-term Scalogram for Acoustic Scene Classification
• Semi-supervised Medical Image Classification with Global Latent Mixing
• A Theory-Driven Self-Labeling Refinement Method for Contrastive Representation Learning
• R-Mixup: Riemannian Mixup for Biological Networks
• Hybrid Composition with IdleBlock: More Efficient Networks for Image Recognition
• MixSearch: Searching for Domain Generalized Medical Image Segmentation Architectures
• S4RL: Surprisingly Simple Self-Supervision for Offline Reinforcement Learning
• Re-labeling ImageNet: from Single to Multi-Labels, from Global to Localized Labels
• DARC: Differentiable ARchitecture Compression
• MixBoost: Synthetic Oversampling with Boosted Mixup for Handling Extreme Imbalance
• Toward Learning Human-aligned Cross-domain Robust Models by Countering Misaligned Features
• S-Mixup: Structural Mixup for Graph Neural Networks
• An Image Patch is a Wave: Phase-Aware Vision MLP
• Distilling Ensemble of Explanations for Weakly-Supervised Pre-Training of Image Segmentation Models
• Augmentation Matters: A Mix-Paste Method for X-Ray Prohibited Item Detection under Noisy Annotations
• MetaBalance: High-Performance Neural Networks for Class-Imbalanced Data
• Challenges for cognitive decoding using deep learning methods
• Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks
• An Ensemble Noise-Robust K-fold Cross-Validation Selection Method for Noisy Labels
• ImageNet-21K Pretraining for the Masses
• GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference
• SamplingAug: On the Importance of Patch Sampling Augmentation for Single Image Super-Resolution
• Deep Bilevel Learning
• PydMobileNet: Improved Version of MobileNets with Pyramid Depthwise Separable Convolution
• Live Face De-Identification in Video
• Multi-Scale High-Resolution Vision Transformer for Semantic Segmentation
• Learning with Hierarchical Complement Objective
• Inner-Imaging Networks: Put Lenses into Convolutional Structure
• A Generalized and Robust Method Towards Practical Gaze Estimation on Smart Phone
• First Order Ambisonics Domain Spatial Augmentation for DNN-based Direction of Arrival Estimation
• CRNNs for Urban Sound Tagging with spatiotemporal context
• Exploring Uncertainty Measures for Image-Caption Embedding-and-Retrieval Task
• Pyramid Adversarial Training Improves ViT Performance
• OnlineAugment: Online Data Augmentation with Less Domain Knowledge
• Deep Learning-Based Strategy for Macromolecules Classification with Imbalanced Data from Cellular Electron Cryotomography
• Latent Domain Learning with Dynamic Residual Adapters
• How Not to Give a FLOP: Combining Regularization and Pruning for Efficient Inference
• A comprehensive study on the prediction reliability of graph neural networks for virtual screening
• BS-NAS: Broadening-and-Shrinking One-Shot NAS with Searchable Numbers of Channels
• Bridge the Domain Gap Between Ultra-wide-field and Traditional Fundus Images via Adversarial Domain Adaptation
• Neural Networks Are More Productive Teachers Than Human Raters: Active Mixup for Data-Efficient Knowledge Distillation from a Blackbox Model
• Attribute Mix: Semantic Data Augmentation for Fine Grained Recognition
• CEKD:Cross Ensemble Knowledge Distillation for Augmented Fine-grained Data
• A Channel Coding Benchmark for Meta-Learning
• Semantically-Conditioned Negative Samples for Efficient Contrastive Learning
• Chest X-ray lung and heart segmentation based on minimal training sets
• Ranking Deep Learning Generalization using Label Variation in Latent Geometry Graphs
• Does Data Augmentation Benefit from Split BatchNorms
• Analyzing Effects of Mixed Sample Data Augmentation on Model Interpretability
• Urban Sound Classification : striving towards a fair comparison
• Center-wise Local Image Mixture For Contrastive Representation Learning
• DeepHIPS: A novel Deep Learning based Hippocampus Subfield Segmentation method
• Federated Multi-Target Domain Adaptation
• An Empirical Study and Analysis on Open-Set Semi-Supervised Learning
• Semi-supervised and Unsupervised Methods for Heart Sounds Classification in Restricted Data Environments
• Revisiting data augmentation for subspace clustering
• A Review and Implementation of Object Detection Models and Optimizations for Real-time Medical Mask Detection during the COVID-19 Pandemic
• Semi-Supervised Learning under General Causal Models
• The Role of Teacher Calibration in Knowledge Distillation
• Downstream-Pretext Domain Knowledge Traceback for Active Learning
• TYolov5: A Temporal Yolov5 Detector Based on Quasi-Recurrent Neural Networks for Real-Time Handgun Detection in Video
• Causality-inspired Single-source Domain Generalization for Medical Image Segmentation
• Learning with Noisy Labels by Efficient Transition Matrix Estimation to Combat Label Miscorrection
• Robustness through Data Augmentation Loss Consistency
• Noisy Annotation Refinement for Object Detection
• Investigate the Essence of Long-Tailed Recognition from a Unified Perspective
• Partial Graph Reasoning for Neural Network Regularization
• HR-NAS: Searching Efficient High-Resolution Neural Architectures with Lightweight Transformers
• Global Wheat Challenge 2020: Analysis of the competition design and winning models
• Go Wider Instead of Deeper
• Recognizing bird species in diverse soundscapes under weak supervision
• Learning Invariant Representation with Consistency and Diversity for Semi-supervised Source Hypothesis Transfer
• Squeeze-Excitation Convolutional Recurrent Neural Networks for Audio-Visual Scene Classification
• GISE-51: A scalable isolated sound events dataset
• Simple Modifications to Improve Tabular Neural Networks
• Exploring the Uncertainty Properties of Neural Networks' Implicit Priors in the Infinite-Width Limit
• How Out-of-Distribution Data Hurts Semi-Supervised Learning
• StackMix: A complementary Mix algorithm
• High-Capacity Expert Binary Networks
• Towards Better Object Detection in Scale Variation with Adaptive Feature Selection
• Data augmentation and image understanding
• Random Shadows and Highlights: A new data augmentation method for extreme lighting conditions
• Uncertainty-guided Model Generalization to Unseen Domains
• PriorityCut: Occlusion-guided Regularization for Warp-based Image Animation
• ROAM: Random Layer Mixup for Semi-Supervised Learning in Medical Imaging
• Counting Out Time: Class Agnostic Video Repetition Counting in the Wild
• Learning Frame Level Attention for Environmental Sound Classification
• F2GAN: Fusing-and-Filling GAN for Few-shot Image Generation
• Learning Temporally Invariant and Localizable Features via Data Augmentation for Video Recognition
• Intelligence plays dice: Stochasticity is essential for machine learning
• Annotation-Free Cardiac Vessel Segmentation via Knowledge Transfer from Retinal Images
• HODGEPODGE: Sound event detection based on ensemble of semi-supervised learning methods
• Safeguarded Dynamic Label Regression for Generalized Noisy Supervision
• Adversarial Transformations for Semi-Supervised Learning
• Learning Multi-instrument Classification with Partial Labels
• Gradually Updated Neural Networks for Large-Scale Image Recognition
• Polyphonic sound event detection based on convolutional recurrent neural networks with semi-supervised loss function for DCASE challenge 2020 task 4
• Improving Polyphonic Sound Event Detection on Multichannel Recordings with the Sørensen-Dice Coefficient Loss and Transfer Learning
• Pseudo-Representation Labeling Semi-Supervised Learning
• Deep Learning Methods for Real-time Detection and Analysis of Wagner Ulcer Classification System
• Learn-able parameter guided Activation Functions
• Locally Adaptive Label Smoothing for Predictive Churn
• Deep Feature Response Discriminative Calibration
• Inference of visual field test performance from OCT volumes using deep learning
• Efficient Method for Categorize Animals in the Wild
• Neural Function Modules with Sparse Arguments: A Dynamic Approach to Integrating Information across Layers
• LIGAR: Lightweight General-purpose Action Recognition
• Semi-supervised Contrastive Learning for Label-efficient Medical Image Segmentation
• Look at here : Utilizing supervision to attend subtle key regions
• InfoScrub: Towards Attribute Privacy by Targeted Obfuscation
• Robustness via Cross-Domain Ensembles
• MUSIQ: Multi-scale Image Quality Transformer
• Negative-ResNet: Noisy Ambulatory Electrocardiogram Signal Classification Scheme
• Multi-stage Optimization based Adversarial Training
• Consistency Regularization with Generative Adversarial Networks for Semi-Supervised Learning
• On the Orthogonality of Knowledge Distillation with Other Techniques: From an Ensemble Perspective
• The Principle of Unchanged Optimality in Reinforcement Learning Generalization
• FixNorm: Dissecting Weight Decay for Training Deep Neural Networks
• PolyViT: Co-training Vision Transformers on Images, Videos and Audio
• Robust Semantic Interpretability: Revisiting Concept Activation Vectors
• An End-to-End Audio Classification System based on Raw Waveforms and Mix-Training Strategy
• Deep Grading based on Collective Artificial Intelligence for AD Diagnosis and Prognosis
• Patch augmentation: Towards efficient decision boundaries for neural networks
• Grammar-Based Grounded Lexicon Learning
• Anatomy Prior Based U-net for Pathology Segmentation with Attention
• A Simple Non-i.i.d. Sampling Approach for Efficient Training and Better Generalization
• Limited Gradient Descent: Learning With Noisy Labels
• Capturing scattered discriminative information using a deep architecture in acoustic scene classification
• Enhancing Mixup-based Semi-Supervised Learning with Explicit Lipschitz Regularization
• SSMD: Semi-Supervised Medical Image Detection with Adaptive Consistency and Heterogeneous Perturbation
• Query-Free Attacks on Industry-Grade Face Recognition Systems under Resource Constraints
• GPT3Mix: Leveraging Large-scale Language Models for Text Augmentation
• k-Mixup Regularization for Deep Learning via Optimal Transport
• Reinforcement Learning with Feedback Graphs
• Churn Reduction via Distillation
• Semi-supervised Domain Adaptation for Semantic Segmentation
• DPNET: Dual-Path Network for Efficient Object Detectioj with Lightweight Self-Attention
• Data Augmentation Imbalance For Imbalanced Attribute Classification
• Data augmentation with Mobius transformations
• Adversarial AutoMixup
• LA3: Efficient Label-Aware AutoAugment
• Augmented Shortcuts for Vision Transformers
• Gi and Pal Scores: Deep Neural Network Generalization Statistics
• Ensemble Learning with Manifold-Based Data Splitting for Noisy Label Correction
• UniMoCo: Unsupervised, Semi-Supervised and Full-Supervised Visual Representation Learning
• TumorCP: A Simple but Effective Object-Level Data Augmentation for Tumor Segmentation
• EDDA: Explanation-driven Data Augmentation to Improve Explanation Faithfulness
• Sample Prior Guided Robust Model Learning to Suppress Noisy Labels
• Unsupervised data augmentation for object detection
• MixRep: Hidden Representation Mixup for Low-Resource Speech Recognition
• Local Patch AutoAugment with Multi-Agent Collaboration
• Electrocardio Panorama: Synthesizing New ECG Views with Self-supervision
• Consistent Posterior Distributions under Vessel-Mixing: A Regularization for Cross-Domain Retinal Artery/Vein Classification
• How Data Augmentation affects Optimization for Linear Regression
• AirMixML: Over-the-Air Data Mixup for Inherently Privacy-Preserving Edge Machine Learning
• Temporally Resolution Decrement: Utilizing the Shape Consistency for Higher Computational Efficiency
• Energy Aligning for Biased Models
• LADA: Look-Ahead Data Acquisition via Augmentation for Active Learning
• Contrastive Mixup: Self- and Semi-Supervised learning for Tabular Domain
• ParamCrop: Parametric Cubic Cropping for Video Contrastive Learning
• Do CNNs Encode Data Augmentations?
• Removing Undesirable Feature Contributions Using Out-of-Distribution Data
• DeepMix: Online Auto Data Augmentation for Robust Visual Object Tracking
• Semixup: In- and Out-of-Manifold Regularization for Deep Semi-Supervised Knee Osteoarthritis Severity Grading from Plain Radiographs
• Discover the Mysteries of the Maya: Selected Contributions from the Machine Learning Challenge & The Discovery Challenge Workshop at ECML PKDD 2021
• TD3Net: A temporal densely connected multi-dilated convolutional network for lipreading
• A Deep Learning Approach for Masking Fetal Gender in Ultrasound Images
• Halo: Learning Semantics-Aware Representations for Cross-Lingual Information Extraction
• A Visual Domain Transfer Learning Approach for Heartbeat Sound Classification
• BasisNet: Two-stage Model Synthesis for Efficient Inference
• COUnty aggRegation mixup AuGmEntation (COURAGE) COVID-19 Prediction
• GroupMixNorm Layer for Learning Fair Models
• Ensemble Model with Batch Spectral Regularization and Data Blending for Cross-Domain Few-Shot Learning with Unlabeled Data
• Towards Robust Pattern Recognition: A Review
• Investigating Group Distributionally Robust Optimization for Deep Imbalanced Learning: A Case Study of Binary Tabular Data Classification
• Deep learning achieves perfect anomaly detection on 108,308 retinal images including unlearned diseases
• MetaAugment: Sample-Aware Data Augmentation Policy Learning
• Principled learning method for Wasserstein distributionally robust optimization with local perturbations
• Transferring Voice Knowledge for Acoustic Event Detection: An Empirical Study
• Bridging the Generalization Gap: Training Robust Models on Confounded Biological Data
• Patch-level Neighborhood Interpolation: A General and Effective Graph-based Regularization Strategy
• Vicinal Feature Statistics Augmentation for Federated 3D Medical Volume Segmentation
• Domain-Aware Augmentations for Unsupervised Online General Continual Learning
• Maximum Entropy Regularization and Chinese Text Recognition
• Context-Gated Convolution
• Boosting Semi-Supervised Face Recognition with Noise Robustness
• Scale Calibrated Training: Improving Generalization of Deep Networks via Scale-Specific Normalization
• Disturbing Target Values for Neural Network Regularization
• Noisy Label Classification using Label Noise Selection with Test-Time Augmentation Cross-Entropy and NoiseMix Learning
• Robust Training Using Natural Transformation
• Multi-dataset Pretraining: A Unified Model for Semantic Segmentation
• Tackling Diverse Minorities in Imbalanced Classification
• Multi-view Feature Augmentation with Adaptive Class Activation Mapping
• Self-supervised GAN Detector
• Ferrograph image classification
• Style Curriculum Learning for Robust Medical Image Segmentation
• Weakly-Supervised Dense Action Anticipation
• Training Classifiers that are Universally Robust to All Label Noise Levels
• Data Augmentation via Mixed Class Interpolation using Cycle-Consistent Generative Adversarial Networks Applied to Cross-Domain Imagery
• Understanding the Disharmony between Weight Normalization Family and Weight Decay: $\epsilon -$shifted $L_2$ Regularizer
• Optimizing Neural Network Scale for ECG Classification
• CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning
• GradFreeBits: Gradient Free Bit Allocation for Dynamic Low Precision Neural Networks
• AMPLIFY:Attention-based Mixup for Performance Improvement and Label Smoothing in Transformer
• HERO: Hessian-Enhanced Robust Optimization for Unifying and Improving Generalization and Quantization Performance
• Enhancing Data-Free Adversarial Distillation with Activation Regularization and Virtual Interpolation
• Synthesising Multi-Modal Minority Samples for Tabular Data
• SIPA: A Simple Framework for Efficient Networks
• Gradient-based Data Augmentation for Semi-Supervised Learning
• Medical Image Segmentation with Limited Supervision: A Review of Deep Network Models
• Object-Aware Cropping for Self-Supervised Learning
• Efficient data augmentation using graph imputation neural networks
• ShuffleBlock: Shuffle to Regularize Deep Convolutional Neural Networks
• Flip Learning: Erase to Segment
• An Enhanced Prohibited Items Recognition Model
• Global and Local Texture Randomization for Synthetic-to-Real Semantic Segmentation
• Calibrating Class Activation Maps for Long-Tailed Visual Recognition
• FlipReID: Closing the Gap between Training and Inference in Person Re-Identification
• MixMix: All You Need for Data-Free Compression Are Feature and Data Mixing
• OpTorch: Optimized deep learning architectures for resource limited environments
• Evaluation of Out-of-Distribution Detection Performance of Self-Supervised Learning in a Controllable Environment
• AI Enlightens Wireless Communication: Analyses, Solutions and Opportunities on CSI Feedback
• Matching the Clinical Reality: Accurate OCT-Based Diagnosis From Few Labels
• A Diversity-Enhanced and Constraints-Relaxed Augmentation for Low-Resource Classification
• Self-supervision of Feature Transformation for Further Improving Supervised Learning
• Hiding Images into Images with Real-world Robustness
• Adversarial Data Encryption
• Graph Partner Neural Networks for Semi-Supervised Learning on Graphs
• Hodge and Podge: Hybrid Supervised Sound Event Detection with Multi-Hot MixMatch and Composition Consistence Training
• Bootstrap Your Object Detector via Mixed Training
• Suppressing Mislabeled Data via Grouping and Self-Attention
• Aggregated Learning: A Deep Learning Framework Based on Information-Bottleneck Vector Quantization
• Soft-Median Choice: An Automatic Feature Smoothing Method for Sound Event Detection
• Elastoformer: Enabling Dynamic Adaptivity via Elastic Model Transformation
• Improving Generalization of Deep Fault Detection Models in the Presence of Mislabeled Data
• LiDAM: Semi-Supervised Learning with Localized Domain Adaptation and Iterative Matching
• Regularizing Recurrent Neural Networks via Sequence Mixup
• Mixup of Feature Maps in a Hidden Layer for Training of Convolutional Neural Network
• White Paper Assistance: A Step Forward Beyond the Shortcut Learning
• Dual Pattern Learning Networks by Empirical Dual Prediction Risk Minimization
• Fractal Dimension Generalization Measure
• Tips and Tricks for Webly-Supervised Fine-Grained Recognition: Learning from the WebFG 2020 Challenge
• How does the Combined Risk Affect the Performance of Unsupervised Domain Adaptation Approaches?
• SDA: Improving Text Generation with Self Data Augmentation
• Improve Global Glomerulosclerosis Classification with Imbalanced Data using CircleMix Augmentation
• DA-Cal: Towards Cross-Domain Calibration in Semantic Segmentation
• Pathology-Aware Generative Adversarial Networks for Medical Image Augmentation
• Know Your Surroundings: Panoramic Multi-Object Tracking by Multimodality Collaboration
• Deep Epidemiological Modeling by Black-box Knowledge Distillation: An Accurate Deep Learning Model for COVID-19
• A novel multiple instance learning framework for COVID-19 severity assessment via data augmentation and self-supervised learning
• Audiovisual transfer learning for audio tagging and sound event detection
• Improving Uncertainty Calibration via Prior Augmented Data
• Fast Camera Image Denoising on Mobile GPUs with Deep Learning, Mobile AI 2021 Challenge: Report
• Towards Efficiently Diversifying Dialogue Generation via Embedding Augmentation
• Geo-Spatiotemporal Features and Shape-Based Prior Knowledge for Fine-grained Imbalanced Data Classification
• Bringing Balance to Hand Shape Classification: Mitigating Data Imbalance Through Generative Models
• Domain-Agnostic Clustering with Self-Distillation
• Research on Optimization Method of Multi-scale Fish Target Fast Detection Network
• Learning Metrics from Mean Teacher: A Supervised Learning Method for Improving the Generalization of Speaker Verification System
• An empirical study of weakly supervised audio tagging embeddings for general audio representations
• SynFace: Face Recognition with Synthetic Data
• Blind Image Decomposition
• Triplet Loss for Knowledge Distillation
• Geometry-Aware Self-Training for Unsupervised Domain Adaptationon Object Point Clouds
• Proxy Experience Replay: Federated Distillation for Distributed Reinforcement Learning
• Learning to Aggregate and Refine Noisy Labels for Visual Sentiment Analysis
• Interlayer and Intralayer Scale Aggregation for Scale-invariant Crowd Counting
• Ripple Attention for Visual Perception with Sub-quadratic Complexity
• Generalization by Recognizing Confusion
• Sequential Feature Filtering Classifier
• Towards Data-Free Domain Generalization
• Multi-Scale Spatially-Asymmetric Recalibration for Image Classification
• Exploring Graph Classification Techniques Under Low Data Constraints: A Comprehensive Study
• Stochastic Region Pooling: Make Attention More Expressive
• Survey on Deep Learning-based Kuzushiji Recognition
• Learning advisor networks for noisy image classification
• Observations on K-image Expansion of Image-Mixing Augmentation for Classification
• Sub-Spectrogram Segmentation for Environmental Sound Classification via Convolutional Recurrent Neural Network and Score Level Fusion
• Dynamic Flows on Curved Space Generated by Labeled Data
• Graph Neural Networks for UnsupervisedDomain Adaptation of Histopathological ImageAnalytics
• Contrastive Vicinal Space for Unsupervised Domain Adaptation
• A Point-Neighborhood Learning Framework for Nasal Endoscope Image Segmentation
• Local Additivity Based Data Augmentation for Semi-supervised NER
• Information based Deep Clustering: An experimental study
• Noisy Feature Mixup
• On Data-centric Myths
• Variance Reduction in Deep Learning: More Momentum is All You Need
• Chinese Herbal Recognition based on Competitive Attentional Fusion of Multi-hierarchies Pyramid Features
• Using mixup as regularization and tuning hyper-parameters for ResNets
• Global Interaction Modelling in Vision Transformer via Super Tokens
• Learning mappings onto regularized latent spaces for biometric authentication
• Unified Multifaceted Feature Learning for Person Re-Identification
• MC-SSL0.0: Towards Multi-Concept Self-Supervised Learning
• Improved Robustness of Vision Transformer via PreLayerNorm in Patch Embedding
• A Generalization Theory based on Independent and Task-Identically Distributed Assumption
• Adversarial Attacks on Cognitive Self-Organizing Networks: The Challenge and the Way Forward
• Signed Input Regularization
• Epsilon Consistent Mixup: Structural Regularization with an Adaptive Consistency-Interpolation Tradeoff
• INN: A Method Identifying Clean-annotated Samples via Consistency Effect in Deep Neural Networks
• Woodscape Fisheye Semantic Segmentation for Autonomous Driving -- CVPR 2021 OmniCV Workshop Challenge
• Semi-supervised learning method based on predefined evenly-distributed class centroids
• Distance Metric-Based Learning with Interpolated Latent Features for Location Classification in Endoscopy Image and Video
• Understanding Invariance via Feedforward Inversion of Discriminatively Trained Classifiers
• Recent Advances in Large Margin Learning
• Searching for an Effective Defender: Benchmarking Defense against Adversarial Word Substitution
• Coded-InvNet for Resilient Prediction Serving Systems