Understanding Neural Networks Through Deep Visualization

Abstract

Recent years have produced great advances in training large, deep neural networks (DNNs), including notable successes in training convolutional neural networks (convnets) to recognize natural images. However, our understanding of how these models work, especially what computations they perform at intermediate layers, has lagged behind. Progress in the field will be further accelerated by the development of better tools for visualizing and interpreting neural nets. We introduce two such tools here. The first is a tool that visualizes the activations produced on each layer of a trained convnet as it processes an image or video (e.g. a live webcam stream). We have found that looking at live activations that change in response to user input helps build valuable intuitions about how convnets work. The second tool enables visualizing features at each layer of a DNN via regularized optimization in image space. Because previous versions of this idea produced less recognizable images, here we introduce several new regularization methods that combine to produce qualitatively clearer, more interpretable visualizations. Both tools are open source and work on a pre-trained convnet with minimal setup.

12 pages. To appear at ICML Deep Learning Workshop 2015

References in corpus (6)

• Improving neural networks by preventing co-adaptation of feature detectors
• Intriguing properties of neural networks
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
• Visualizing and Understanding Convolutional Networks
• Pylearn2: a machine learning research library

Cited by in corpus (364)

• Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks
• Axiomatic Attribution for Deep Networks
• A Survey on Explainable Artificial Intelligence (XAI): Towards Medical XAI
• DeepXplore: Automated Whitebox Testing of Deep Learning Systems
• Perceptual Losses for Real-Time Style Transfer and Super-Resolution
• WaveletKernelNet: An Interpretable Deep Neural Network for Industrial Intelligent Diagnosis
• Towards Explainable Artificial Intelligence
• This Looks Like That: Deep Learning for Interpretable Image Recognition
• AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery
• Visualizing Deep Convolutional Neural Networks Using Natural Pre-Images
• Deep image mining for diabetic retinopathy screening
• Boosting the accuracy of multi-spectral image pan-sharpening by learning a deep residual network
• Deep Clustering for Unsupervised Learning of Visual Features
• Predicting Process Behaviour using Deep Learning
• Visualizing Deep Neural Network Decisions: Prediction Difference Analysis
• Deep Anomaly Detection Using Geometric Transformations
• DeepSeeNet: A deep learning model for automated classification of patient-based age-related macular degeneration severity from color fundus photographs
• Robustness May Be at Odds with Accuracy
• CNN Explainer: Learning Convolutional Neural Networks with Interactive Visualization
• Recent Advances in Convolutional Neural Networks
• Physically Interpretable Neural Networks for the Geosciences: Applications to Earth System Variability
• Layer-Wise Relevance Propagation for Explaining Deep Neural Network Decisions in MRI-Based Alzheimer's Disease Classification
• What Clinicians Want: Contextualizing Explainable Machine Learning for Clinical End Use
• Synthesizing the preferred inputs for neurons in neural networks via deep generator networks
• A Survey of Privacy Attacks in Machine Learning
• Manifold: A Model-Agnostic Framework for Interpretation and Diagnosis of Machine Learning Models
• Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI
• GAN Lab: Understanding Complex Deep Generative Models using Interactive Visual Experimentation
• Do Convolutional Neural Networks Learn Class Hierarchy?
• Looking Beyond Appearances: Synthetic Training Data for Deep CNNs in Re-identification
• High-Resolution Breast Cancer Screening with Multi-View Deep Convolutional Neural Networks
• DarkneTZ: Towards Model Privacy at the Edge using Trusted Execution Environments
• The State of the Art in Enhancing Trust in Machine Learning Models with the Use of Visualizations
• Testing Deep Neural Networks
• Multifaceted Feature Visualization: Uncovering the Different Types of Features Learned By Each Neuron in Deep Neural Networks
• Smooth Grad-CAM++: An Enhanced Inference Level Visualization Technique for Deep Convolutional Neural Network Models
• Deep Learning and Its Applications to Machine Health Monitoring: A Survey
• Automatic Sleep Staging of EEG Signals: Recent Development, Challenges, and Future Directions
• Spinal cord gray matter segmentation using deep dilated convolutions
• A Fully Convolutional Network for Semantic Labeling of 3D Point Clouds
• Galaxy Morphology Classification with Deep Convolutional Neural Networks
• Generating and designing DNA with deep generative models
• One Model to Synthesize Them All: Multi-contrast Multi-scale Transformer for Missing Data Imputation
• Visualizing Convolutional Neural Network Protein-Ligand Scoring
• TSViz: Demystification of Deep Learning Models for Time-Series Analysis
• What Do We Understand About Convolutional Networks?
• Machine learning with neural networks
• Universal Adversarial Examples in Remote Sensing: Methodology and Benchmark
• Deep Neural Network for Real-Time Autonomous Indoor Navigation
• Hierarchical interpretations for neural network predictions
• Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting
• Attribute2Image: Conditional Image Generation from Visual Attributes
• Generalisation and Sharing in Triplet Convnets for Sketch based Visual Search
• Explainable Artificial Intelligence: a Systematic Review
• High Accuracy Tumor Diagnoses and Benchmarking of Hematoxylin and Eosin Stained Prostate Core Biopsy Images Generated by Explainable Deep Neural Networks
• Gradients of Counterfactuals
• Charting the Sociotechnical Gap in Explainable AI: A Framework to Address the Gap in XAI
• Explaining Vulnerabilities to Adversarial Machine Learning through Visual Analytics
• Neural Best-Buddies: Sparse Cross-Domain Correspondence
• How Important Is a Neuron?
• Adversarial Robustness as a Prior for Learned Representations
• Accurate Virus Identification with Interpretable Raman Signatures by Machine Learning
• Filter Pruning by Switching to Neighboring CNNs with Good Attributes
• Cognitive Psychology for Deep Neural Networks: A Shape Bias Case Study
• Improving Variational Autoencoder with Deep Feature Consistent and Generative Adversarial Training
• Deep Motif: Visualizing Genomic Sequence Classifications
• A Survey of Deep Learning for Scientific Discovery
• Insights on representational similarity in neural networks with canonical correlation
• Lets keep it simple, Using simple architectures to outperform deeper and more complex architectures
• Spatial Correlation and Value Prediction in Convolutional Neural Networks
• A Survey of Safety and Trustworthiness of Deep Neural Networks: Verification, Testing, Adversarial Attack and Defence, and Interpretability
• A Taxonomy and Library for Visualizing Learned Features in Convolutional Neural Networks
• A Hierarchy of Limitations in Machine Learning
• Beyond saliency: understanding convolutional neural networks from saliency prediction on layer-wise relevance propagation
• AI-GAs: AI-generating algorithms, an alternate paradigm for producing general artificial intelligence
• How convolutional neural network see the world - A survey of convolutional neural network visualization methods
• Using Visual Analytics to Interpret Predictive Machine Learning Models
• On Interpretability of Artificial Neural Networks: A Survey
• VisualBackProp: efficient visualization of CNNs
• CANet: Class-Agnostic Segmentation Networks with Iterative Refinement and Attentive Few-Shot Learning
• Understanding trained CNNs by indexing neuron selectivity
• Automated Rationale Generation: A Technique for Explainable AI and its Effects on Human Perceptions
• Are we done with object recognition? The iCub robot's perspective
• Dank or Not? -- Analyzing and Predicting the Popularity of Memes on Reddit
• A Selective Overview of Deep Learning
• Image Companding and Inverse Halftoning using Deep Convolutional Neural Networks
• Fully Convolutional Networks for Chip-wise Defect Detection Employing Photoluminescence Images
• Saccader: Improving Accuracy of Hard Attention Models for Vision
• Visual Explanation by Interpretation: Improving Visual Feedback Capabilities of Deep Neural Networks
• Unsupervised learning of object semantic parts from internal states of CNNs by population encoding
• Deep speech inpainting of time-frequency masks
• A Comprehensive Survey of Machine Learning Applied to Radar Signal Processing
• Explaining First Impressions: Modeling, Recognizing, and Explaining Apparent Personality from Videos
• An Analysis of Pre-Training on Object Detection
• A Survey on Deep Learning of Small Sample in Biomedical Image Analysis
• Neural Network Attributions: A Causal Perspective
• Generative Counterfactual Introspection for Explainable Deep Learning
• Detecting Statistical Interactions from Neural Network Weights
• Concurrent Activity Recognition with Multimodal CNN-LSTM Structure
• A Survey on Neural Network Interpretability
• Explainable Machine Learning with Prior Knowledge: An Overview
• Towards falsifiable interpretability research
• Rice Classification Using Spatio-Spectral Deep Convolutional Neural Network
• MAMMO: A Deep Learning Solution for Facilitating Radiologist-Machine Collaboration in Breast Cancer Diagnosis
• Detecting gravitational lenses using machine learning: exploring interpretability and sensitivity to rare lensing configurations
• A Study of the Generalizability of Self-Supervised Representations
• Opening the black box of neural nets: case studies in stop/top discrimination
• When and How to Fool Explainable Models (and Humans) with Adversarial Examples
• Regional Multi-scale Approach for Visually Pleasing Explanations of Deep Neural Networks
• CNNComparator: Comparative Analytics of Convolutional Neural Networks
• Interpretability Beyond Classification Output: Semantic Bottleneck Networks
• AI Research Considerations for Human Existential Safety (ARCHES)
• Deep Feature Consistent Deep Image Transformations: Downscaling, Decolorization and HDR Tone Mapping
• Disentangling Image Distortions in Deep Feature Space
• A New Method to Visualize Deep Neural Networks
• Adversarial Neural Network Inversion via Auxiliary Knowledge Alignment
• Interpreting Deep Learning Models in Natural Language Processing: A Review
• Greedy Layerwise Learning Can Scale to ImageNet
• The RFML Ecosystem: A Look at the Unique Challenges of Applying Deep Learning to Radio Frequency Applications
• Efficient Gender Classification Using a Deep LDA-Pruned Net
• Saliency Prediction in the Deep Learning Era: Successes, Limitations, and Future Challenges
• Meta-DETR: Image-Level Few-Shot Object Detection with Inter-Class Correlation Exploitation
• Automatic Photo Orientation Detection with Convolutional Neural Networks
• Relevance Prediction from Eye-movements Using Semi-interpretable Convolutional Neural Networks
• A semi-supervised Machine Learning search for never-seen Gravitational-Wave sources
• Mechanisms of a Convolutional Neural Network for Learning Three-dimensional Unsteady Wake Flow
• What does it mean to understand a neural network?
• Understanding Early Word Learning in Situated Artificial Agents
• Exploring the Function Space of Deep-Learning Machines
• Explaining Convolutional Neural Networks using Softmax Gradient Layer-wise Relevance Propagation
• Towards Principled Design of Deep Convolutional Networks: Introducing SimpNet
• Distilling Critical Paths in Convolutional Neural Networks
• Adapting Object Detectors with Conditional Domain Normalization
• How do Mixture Density RNNs Predict the Future?
• Rationalization: A Neural Machine Translation Approach to Generating Natural Language Explanations
• Selectivity considered harmful: evaluating the causal impact of class selectivity in DNNs
• A Century of Portraits: A Visual Historical Record of American High School Yearbooks
• Deep Nets: What have they ever done for Vision?
• Precomputed Real-Time Texture Synthesis with Markovian Generative Adversarial Networks
• CRNet: Cross-Reference Networks for Few-Shot Segmentation
• Comparing Rule-Based and Deep Learning Models for Patient Phenotyping
• How Much Can I Trust You? -- Quantifying Uncertainties in Explaining Neural Networks
• Unsupervised Learning of Neural Networks to Explain Neural Networks
• Leveraging Adaptive Color Augmentation in Convolutional Neural Networks for Deep Skin Lesion Segmentation
• DUP-Net: Denoiser and Upsampler Network for 3D Adversarial Point Clouds Defense
• Learning Discriminative Motion Features Through Detection
• Evaluating the visualization of what a Deep Neural Network has learned
• Modeling Spatial and Temporal Cues for Multi-label Facial Action Unit Detection
• LCA: Loss Change Allocation for Neural Network Training
• Learning Temporal Pose Estimation from Sparsely-Labeled Videos
• Representation of linguistic form and function in recurrent neural networks
• Confidential Inference via Ternary Model Partitioning
• Explaining Knowledge Distillation by Quantifying the Knowledge
• Deep Learning Development Environment in Virtual Reality
• Explaining Bayesian Neural Networks
• Deeply Explain CNN via Hierarchical Decomposition
• ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models
• Functional connectivity patterns of autism spectrum disorder identified by deep feature learning
• Exact and Consistent Interpretation for Piecewise Linear Neural Networks: A Closed Form Solution
• Deep Feature Consistent Variational Autoencoder
• Automated detection of oral pre-cancerous tongue lesions using deep learning for early diagnosis of oral cavity cancer
• Mediators in Determining what Processing BERT Performs First
• Activation Landscapes as a Topological Summary of Neural Network Performance
• Functionality-Oriented Convolutional Filter Pruning
• Learning to synthesize: splitting and recombining low and high spatial frequencies for image recovery
• What Catches the Eye? Visualizing and Understanding Deep Saliency Models
• Interpretable Deep Convolutional Neural Networks via Meta-learning
• Parkinson's Disease Assessment from a Wrist-Worn Wearable Sensor in Free-Living Conditions: Deep Ensemble Learning and Visualization
• The Intriguing Properties of Model Explanations
• Improving the Interpretability of Deep Neural Networks with Knowledge Distillation
• Statistics of Visual Responses to Object Stimuli from Primate AIT Neurons to DNN Neurons
• Teaching Meaningful Explanations
• Brain Biomarker Interpretation in ASD Using Deep Learning and fMRI
• Multimodal Transfer: A Hierarchical Deep Convolutional Neural Network for Fast Artistic Style Transfer
• Advances in Deep Learning for Hyperspectral Image Analysis--Addressing Challenges Arising in Practical Imaging Scenarios
• AI Blue Book: Vehicle Price Prediction using Visual Features
• Generalizing multistain immunohistochemistry tissue segmentation using one-shot color deconvolution deep neural networks
• Face-to-Parameter Translation for Game Character Auto-Creation
• Plug-and-Play: Improve Depth Estimation via Sparse Data Propagation
• Visualizing and Understanding Sum-Product Networks
• Self-Supervised Feature Learning by Learning to Spot Artifacts
• Relating Input Concepts to Convolutional Neural Network Decisions
• Interpretable Discovery in Large Image Data Sets
• Enhancing the Extraction of Interpretable Information for Ischemic Stroke Imaging from Deep Neural Networks
• DeepFeat: A Bottom Up and Top Down Saliency Model Based on Deep Features of Convolutional Neural Nets
• Towards Better Analysis of Deep Convolutional Neural Networks
• Dynamic Origin-Destination Matrix Prediction with Line Graph Neural Networks and Kalman Filter
• What evidence does deep learning model use to classify Skin Lesions?
• Dissecting Catastrophic Forgetting in Continual Learning by Deep Visualization
• NaturalInversion: Data-Free Image Synthesis Improving Real-World Consistency
• A Game-Theoretic Taxonomy of Visual Concepts in DNNs
• Perceptual underwater image enhancement with deep learning and physical priors
• Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles
• Examining the Benefits of Capsule Neural Networks
• Towards a Unified Evaluation of Explanation Methods without Ground Truth
• Why do These Match? Explaining the Behavior of Image Similarity Models
• GAN-based Generation and Automatic Selection of Explanations for Neural Networks
• Complementary reinforcement learning towards explainable agents
• SISC: End-to-end Interpretable Discovery Radiomics-Driven Lung Cancer Prediction via Stacked Interpretable Sequencing Cells
• Classifying the classifier: dissecting the weight space of neural networks
• Neural Network Memorization Dissection
• Multi-level Multimodal Common Semantic Space for Image-Phrase Grounding
• Demystifying Neural Network Filter Pruning
• Learning with Interpretable Structure from Gated RNN
• Respond-CAM: Analyzing Deep Models for 3D Imaging Data by Visualizations
• Egocentric Basketball Motion Planning from a Single First-Person Image
• Autoencoder based image compression: can the learning be quantization independent?
• Deep saliency: What is learnt by a deep network about saliency?
• Do Vision Models Encode Object-Level Semantic Relatedness? A Cognitive Psychology-Inspired Benchmark
• Picasso: A Modular Framework for Visualizing the Learning Process of Neural Network Image Classifiers
• Multi-task learning with deep model based reinforcement learning
• Happy Travelers Take Big Pictures: A Psychological Study with Machine Learning and Big Data
• Explainable Multivariate Time Series Classification: A Deep Neural Network Which Learns To Attend To Important Variables As Well As Informative Time Intervals
• Deep segmentation networks predict survival of non-small cell lung cancer
• Convolutional Neural Networks Do Work with Pre-Defined Filters
• Interpreting Shared Deep Learning Models via Explicable Boundary Trees
• Food Ingredients Recognition through Multi-label Learning
• IMAGINE: Image Synthesis by Image-Guided Model Inversion
• MonoNet: Towards Interpretable Models by Learning Monotonic Features
• Exemplary Natural Images Explain CNN Activations Better than State-of-the-Art Feature Visualization
• A Survey on Understanding, Visualizations, and Explanation of Deep Neural Networks
• Human-Expert-Level Brain Tumor Detection Using Deep Learning with Data Distillation and Augmentation
• Explainable Deep Learning for Uncovering Actionable Scientific Insights for Materials Discovery and Design
• NEUZZ: Efficient Fuzzing with Neural Program Smoothing
• HWNet v2: An Efficient Word Image Representation for Handwritten Documents
• Can We Faithfully Represent Masked States to Compute Shapley Values on a DNN?
• Let's Agree to Agree: Neural Networks Share Classification Order on Real Datasets
• Interactive Visualization for Debugging RL
• Server, server in the cloud. Who is the fairest in the crowd?
• Boundary Aware U-Net for Glacier Segmentation
• Darker than Black-Box: Face Reconstruction from Similarity Queries
• TSInsight: A local-global attribution framework for interpretability in time-series data
• Domain-Level Explainability -- A Challenge for Creating Trust in Superhuman AI Strategies
• HypoML: Visual Analysis for Hypothesis-based Evaluation of Machine Learning Models
• What Does a TextCNN Learn?
• The Representation Theory of Neural Networks
• Answering Questions about Data Visualizations using Efficient Bimodal Fusion
• Detecting Colorized Images via Convolutional Neural Networks: Toward High Accuracy and Good Generalization
• Visualizing Deep Neural Networks for Speech Recognition with Learned Topographic Filter Maps
• Understanding Deep Architectures by Visual Summaries
• CNNs for Surveillance Footage Scene Classification
• GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference
• Normalizing Flow based Hidden Markov Models for Classification of Speech Phones with Explainability
• Fine-grained visual recognition with salient feature detection
• CAFENet: Class-Agnostic Few-Shot Edge Detection Network
• Explaining AlphaGo: Interpreting Contextual Effects in Neural Networks
• Defining and Quantifying the Emergence of Sparse Concepts in DNNs
• Attacking Convolutional Neural Network using Differential Evolution
• Deep Net Triage: Analyzing the Importance of Network Layers via Structural Compression
• Deep Neural Models for color discrimination and color constancy
• Leveraging Sparse Linear Layers for Debuggable Deep Networks
• Top-down Neural Attention by Excitation Backprop
• Efficient Image Evidence Analysis of CNN Classification Results
• A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution
• A Topological Framework for Deep Learning
• On the Behavior of Convolutional Nets for Feature Extraction
• Computing Linear Restrictions of Neural Networks
• XDeep: An Interpretation Tool for Deep Neural Networks
• Path Analysis for Effective Fault Localization in Deep Neural Networks
• Gradient-Coherent Strong Regularization for Deep Neural Networks
• Visualizing Image Content to Explain Novel Image Discovery
• Fairness in Deep Learning: A Computational Perspective
• NASirt: AutoML based learning with instance-level complexity information
• Interpretable Neural Network Decoupling
• DeepEverest: Accelerating Declarative Top-K Queries for Deep Neural Network Interpretation
• Efficient Action Recognition Using Confidence Distillation
• Fast gradient-free activation maximization for neurons in spiking neural networks
• Model Interpretability and Rationale Extraction by Input Mask Optimization
• DeepChrome: Deep-learning for predicting gene expression from histone modifications
• Towards Explanation of DNN-based Prediction with Guided Feature Inversion
• SoPa: Bridging CNNs, RNNs, and Weighted Finite-State Machines
• Visualizing MuZero Models
• Intraclass clustering: an implicit learning ability that regularizes DNNs
• Actionable Attribution Maps for Scientific Machine Learning
• S4Net: Single Stage Salient-Instance Segmentation
• Transparent Classification with Multilayer Logical Perceptrons and Random Binarization
• Randomness in Deconvolutional Networks for Visual Representation
• Automated Testing for Deep Learning Systems with Differential Behavior Criteria
• Sampling for Deep Learning Model Diagnosis (Technical Report)
• Neuron ranking -- an informed way to condense convolutional neural networks architecture
• Convolutional Neural Network on Three Orthogonal Planes for Dynamic Texture Classification
• Scalable Topological Data Analysis and Visualization for Evaluating Data-Driven Models in Scientific Applications
• Learning Spatiotemporal Features for Infrared Action Recognition with 3D Convolutional Neural Networks
• Modeling Latent Attention Within Neural Networks
• Deep Epitome for Unravelling Generalized Hamming Network: A Fuzzy Logic Interpretation of Deep Learning
• Sampling the "Inverse Set" of a Neuron: An Approach to Understanding Neural Nets
• TopoAct: Visually Exploring the Shape of Activations in Deep Learning
• Pretraining Image Encoders without Reconstruction via Feature Prediction Loss
• Channel Interaction Networks for Fine-Grained Image Categorization
• Interpreting and Disentangling Feature Components of Various Complexity from DNNs
• Fed2: Feature-Aligned Federated Learning
• Optimising the Input Image to Improve Visual Relationship Detection
• Low Photon Budget Phase Retrieval with Perceptual Loss Trained Deep Neural Networks
• ReNN: Rule-embedded Neural Networks
• Out of the Black Box: Properties of deep neural networks and their applications
• Real Time Visual Tracking using Spatial-Aware Temporal Aggregation Network
• LAMVI-2: A Visual Tool for Comparing and Tuning Word Embedding Models
• The Uncanny Similarity of Recurrence and Depth
• Convolutional Autoencoders for Lossy Light Field Compression
• Rapid Identification of X-ray Diffraction Spectra Based on Very Limited Data by Interpretable Convolutional Neural Networks
• Explaining How Deep Neural Networks Forget by Deep Visualization
• Iterative and Adaptive Sampling with Spatial Attention for Black-Box Model Explanations
• Dirichlet Pruning for Neural Network Compression
• Understanding Patch-Based Learning by Explaining Predictions
• Every Filter Extracts A Specific Texture In Convolutional Neural Networks
• Generalized Constraints as A New Mathematical Problem in Artificial Intelligence: A Review and Perspective
• Visual Understanding of Multiple Attributes Learning Model of X-Ray Scattering Images
• On The Distribution of Penultimate Activations of Classification Networks
• Fuzzy Logic Interpretation of Quadratic Networks
• Why Convolutional Networks Learn Oriented Bandpass Filters: Theory and Empirical Support
• Illuminated Decision Trees with Lucid
• GAM: Explainable Visual Similarity and Classification via Gradient Activation Maps
• Automatic Attribute Discovery with Neural Activations
• Exploring Frequency Domain Interpretation of Convolutional Neural Networks
• Understanding of Kernels in CNN Models by Suppressing Irrelevant Visual Features in Images
• Assembling Semantically-Disentangled Representations for Predictive-Generative Models via Adaptation from Synthetic Domain
• Exploring the Deep Feature Space of a Cell Classification Neural Network
• Soft Sensing Model Visualization: Fine-tuning Neural Network from What Model Learned
• Inverting and Understanding Object Detectors
• Deep Cross Modal Learning for Caricature Verification and Identification(CaVINet)
• Carrying out CNN Channel Pruning in a White Box
• A Framework to Learn with Interpretation
• Associating Grasp Configurations with Hierarchical Features in Convolutional Neural Networks
• Inference Graphs for CNN Interpretation
• CvS: Classification via Segmentation For Small Datasets
• In situ TensorView: In situ Visualization of Convolutional Neural Networks
• Efficient Modelling Across Time of Human Actions and Interactions
• Statistical model-based evaluation of neural networks
• Adversarial-Playground: A Visualization Suite for Adversarial Sample Generation
• Semantic Network Interpretation
• Topological Deep Learning: Classification Neural Networks
• Interpreting deep urban sound classification using Layer-wise Relevance Propagation
• On the Selective and Invariant Representation of DCNN for High-Resolution Remote Sensing Image Recognition
• Explainable Adversarial Attacks in Deep Neural Networks Using Activation Profiles
• Understanding and Improving Kernel Local Descriptors
• A Hybrid Framework for Matching Printing Design Files to Product Photos
• Identifying Layers Susceptible to Adversarial Attacks
• Staging Epileptogenesis with Deep Neural Networks
• Are there any 'object detectors' in the hidden layers of CNNs trained to identify objects or scenes?
• Object Specific Deep Learning Feature and Its Application to Face Detection
• Learning to Play Pong using Policy Gradient Learning
• Understanding learned CNN features through Filter Decoding with Substitution
• Model Reduction of Shallow CNN Model for Reliable Deployment of Information Extraction from Medical Reports
• Attack to Fool and Explain Deep Networks
• Interactive Visual Study of Multiple Attributes Learning Model of X-Ray Scattering Images
• Explaining decision of model from its prediction
• Hierarchical Neural Representation of Dreamed Objects Revealed by Brain Decoding with Deep Neural Network Features
• An Overview of Computational Approaches for Interpretation Analysis
• Detecting Deep Neural Network Defects with Data Flow Analysis
• IFBiD: Inference-Free Bias Detection
• Assisting human experts in the interpretation of their visual process: A case study on assessing copper surface adhesive potency
• Understanding Convolutional Neural Networks with A Mathematical Model
• GRIm-RePR: Prioritising Generating Important Features for Pseudo-Rehearsal
• Capturing Localized Image Artifacts through a CNN-based Hyper-image Representation
• A Deep Convolutional Network for Seismic Shot-Gather Image Quality Classification
• Leaky ReLUs That Differ in Forward and Backward Pass Facilitate Activation Maximization in Deep Neural Networks
• Analyzing Learned Convnet Features with Dirichlet Process Gaussian Mixture Models
• Learning Identity-Preserving Transformations on Data Manifolds
• Gradient-Adjusted Neuron Activation Profiles for Comprehensive Introspection of Convolutional Speech Recognition Models
• How to Explain Neural Networks: an Approximation Perspective
• Training Deep Neural Networks via Optimization Over Graphs
• What's the relationship between CNNs and communication systems?
• Det-CGD: Compressed Gradient Descent with Matrix Stepsizes for Non-Convex Optimization
• LOTS about Attacking Deep Features