Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms

Abstract

We present Fashion-MNIST, a new dataset comprising of 28x28 grayscale images of 70,000 fashion products from 10 categories, with 7,000 images per category. The training set has 60,000 images and the test set has 10,000 images. Fashion-MNIST is intended to serve as a direct drop-in replacement for the original MNIST dataset for benchmarking machine learning algorithms, as it shares the same image size, data format and the structure of training and testing splits. The dataset is freely available at https://github.com/zalandoresearch/fashion-mnist

Dataset is freely available at https://github.com/zalandoresearch/fashion-mnist Benchmark is available at http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/

References in corpus (1)

• EMNIST: an extension of MNIST to handwritten letters

Cited by in corpus (83)

• Random Erasing Data Augmentation
• Agnostic Federated Learning
• Peer-to-peer Federated Learning on Graphs
• Performance-Efficiency Trade-off of Low-Precision Numerical Formats in Deep Neural Networks
• Superposition of many models into one
• Capsule Networks with Max-Min Normalization
• A framework for the extraction of Deep Neural Networks by leveraging public data
• Measurements of Three-Level Hierarchical Structure in the Outliers in the Spectrum of Deepnet Hessians
• Bridging Adversarial Robustness and Gradient Interpretability
• Decentralized Bayesian Learning over Graphs
• Learning Generative Models across Incomparable Spaces
• XNAS: Neural Architecture Search with Expert Advice
• CapsAttacks: Robust and Imperceptible Adversarial Attacks on Capsule Networks
• Semi-Supervised Learning by Label Gradient Alignment
• On the Vulnerability of Capsule Networks to Adversarial Attacks
• Latent Translation: Crossing Modalities by Bridging Generative Models
• Evaluating the Robustness of Nearest Neighbor Classifiers: A Primal-Dual Perspective
• Embedded hyper-parameter tuning by Simulated Annealing
• Detecting Adversarial Examples and Other Misclassifications in Neural Networks by Introspection
• Bayesian Learning of Neural Network Architectures
• PuVAE: A Variational Autoencoder to Purify Adversarial Examples
• Deep Density-based Image Clustering
• State-Regularized Recurrent Neural Networks
• Fast Inference in Capsule Networks Using Accumulated Routing Coefficients
• On the Necessity and Effectiveness of Learning the Prior of Variational Auto-Encoder
• Fix Your Features: Stationary and Maximally Discriminative Embeddings using Regular Polytope (Fixed Classifier) Networks
• SwiftNet: Using Graph Propagation as Meta-knowledge to Search Highly Representative Neural Architectures
• Towards Characterizing and Limiting Information Exposure in DNN Layers
• Unsupervised Multi-Domain Image Translation with Domain-Specific Encoders/Decoders
• Identifying Classes Susceptible to Adversarial Attacks
• Better the Devil you Know: An Analysis of Evasion Attacks using Out-of-Distribution Adversarial Examples
• Clustering-Oriented Representation Learning with Attractive-Repulsive Loss
• NeuNetS: An Automated Synthesis Engine for Neural Network Design
• AutoAssist: A Framework to Accelerate Training of Deep Neural Networks
• The functional role of cue-driven feature-based feedback in object recognition
• DeepOBS: A Deep Learning Optimizer Benchmark Suite
• Inductive Transfer for Neural Architecture Optimization
• Wasserstein-Wasserstein Auto-Encoders
• Unsupervised Data Imputation via Variational Inference of Deep Subspaces
• Deep Neural Architecture Search with Deep Graph Bayesian Optimization
• Prediction is very hard, especially about conversion. Predicting user purchases from clickstream data in fashion e-commerce
• Deep Generative Learning via Variational Gradient Flow
• Training CNNs with Selective Allocation of Channels
• Cross-modal Variational Auto-encoder with Distributed Latent Spaces and Associators
• Facilitating Bayesian Continual Learning by Natural Gradients and Stein Gradients
• Combining Model and Parameter Uncertainty in Bayesian Neural Networks
• Reconstruction and Membership Inference Attacks against Generative Models
• Active Generative Adversarial Network for Image Classification
• Training on test data: Removing near duplicates in Fashion-MNIST
• Divergence Triangle for Joint Training of Generator Model, Energy-based Model, and Inference Model
• k-GANs: Ensemble of Generative Models with Semi-Discrete Optimal Transport
• Dueling Decoders: Regularizing Variational Autoencoder Latent Spaces
• Probabilistic Discriminative Learning with Layered Graphical Models
• Incremental Concept Learning via Online Generative Memory Recall
• Prior Activation Distribution (PAD): A Versatile Representation to Utilize DNN Hidden Units
• A Computationally Efficient Method for Defending Adversarial Deep Learning Attacks
• Unsupervised Domain Adaptation using Deep Networks with Cross-Grafted Stacks
• DeepSwarm: Optimising Convolutional Neural Networks using Swarm Intelligence
• Learning Generative Models of Structured Signals from Their Superposition Using GANs with Application to Denoising and Demixing
• Hierarchical Annotation of Images with Two-Alternative-Forced-Choice Metric Learning
• Context-modulation of hippocampal dynamics and deep convolutional networks
• Mapper Comparison with Wasserstein Metrics
• On Transformations in Stochastic Gradient MCMC
• Venn GAN: Discovering Commonalities and Particularities of Multiple Distributions
• Revisiting hard thresholding for DNN pruning
• Deep Supervised Hashing leveraging Quadratic Spherical Mutual Information for Content-based Image Retrieval
• Novel Uncertainty Framework for Deep Learning Ensembles
• Rethinking Layer-wise Feature Amounts in Convolutional Neural Network Architectures
• Spatially-Coupled Neural Network Architectures
• Stochastic Sparse Learning with Momentum Adaptation for Imprecise Memristor Networks
• neuralRank: Searching and ranking ANN-based model repositories
• Probabilistic Decoupling of Labels in Classification
• DaiMoN: A Decentralized Artificial Intelligence Model Network
• Widely Linear Kernels for Complex-Valued Kernel Activation Functions
• CheckNet: Secure Inference on Untrusted Devices
• Stochastic Doubly Robust Gradient
• Generative Adversarial Networks with Inverse Transformation Unit
• Part-based approximations for morphological operators using asymmetric auto-encoders
• Virtual Conditional Generative Adversarial Networks
• Interpretable BoW Networks for Adversarial Example Detection
• A Trainable Multiplication Layer for Auto-correlation and Co-occurrence Extraction
• Stochastic Computing for Hardware Implementation of Binarized Neural Networks
• Generating Adversarial Examples With Conditional Generative Adversarial Net