Repository files navigation

This repository contains short summaries of some machine learning papers.

• UNSUPERVISED LEARNINGECCV 2018Deep Clustering for Unsupervised Learning of Visual Features
• OBJECT DETECTIONPOINT CLOUDSELF-DRIVING CARSECCV 2018Deep Continuous Fusion for Multi-Sensor 3D Object Detection
• AUDIOSOUND SOURCE LOCALIZATIONACTION RECOGNITIONSOUND SOURCE SEPARATIONSELF-SUPERVISEDECCV 2018Audio-Visual Scene Analysis with Self-Supervised Multisensory Features
• UNCERTAINTYECCV 2018Towards Realistic Predictors
• OBJECT DETECTIONECCV 2018Acquisition of Localization Confidence for Accurate Object Detection
• OBJECT DETECTIONECCV 2018CornerNet: Detecting Objects as Paired Keypoints
• NORMALIZATIONECCV 2018Group Normalization
• ARCHITECTURESATTENTIONECCV 2018Convolutional Networks with Adaptive Inference Graphs

• ARCHITECTURESATTENTIONSpatial Transformer Networks (thanks,alexobednikov)

• LOSS FUNCTIONSRECOGNITIONWorking hard to know your neighbor’s margins: Local descriptor learning loss (thanks,alexobednikov)

• FACE RECOGNITIONFACESNeural Aggregation Network for Video Face Recognition (thanks,alexobednikov)

• Critical Learning Periods in Deep Neural Networks
• GANSELF-DRIVING CARSHigh-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs
• SELF-DRIVING CARSComputer Vision for Autonomous Vehicles: Problems, Datasets and State-of-the-Art

• GANProgressive Growing of GANs for Improved Quality, Stability, and Variation

• SELF-DRIVING CARSSystematic Testing of Convolutional Neural Networks for Autonomous Driving
• SELF-DRIVING CARSSEGMENTATIONFast Scene Understanding for Autonomous Driving
• SELF-DRIVING CARSArguing Machines: Perception-Control System Redundancy and Edge Case Discovery in Real-World Autonomous Driving
• SELF-DRIVING CARSGANREINFORCEMENTVirtual to Real Reinforcement Learning for Autonomous Driving
• SELF-DRIVING CARSEnd to End Learning for Self-Driving Cars

• Snapshot Ensembles: Train 1, get M for free
• Image Crowd Counting Using Convolutional Neural Network and Markov Random Field
• REINFORCEMENTRainbow: Combining Improvements in Deep Reinforcement Learning
• REINFORCEMENTLearning to Navigate in Complex Environments
• GANUnsupervised Image-to-Image Translation Networks
• RNNDilated Recurrent Neural Networks
• OBJECT DETECTIONTRACKINGDetect to Track and Track to Detect
• ARCHITECTURESDilated Residual Networks

• OBJECT DETECTIONFeature Pyramid Networks for Object Detection
• OBJECT DETECTIONSSD: Single Shot MultiBox Detector
• OBJECT DETECTIONEFFICIENT NETWORKSMobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• OBJECT DETECTIONMask R-CNN

• FACESMulti-view Face Detection Using Deep Convolutional Neural Networks (aka DDFD) (thanks,arnaldog12)

• GANOn the Effects of Batch and Weight Normalization in Generative Adversarial Networks
• GANBEGAN
• GANStackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networks
• ACTIVATION FUNCTIONSSelf-Normalizing Neural Networks
• GANWasserstein GAN (aka WGAN)

• OBJECT DETECTIONYOLO9000: Better, Faster, Stronger (aka YOLOv2)
• OBJECT DETECTIONYou Only Look Once: Unified, Real-Time Object Detection (aka YOLO)
• OBJECT DETECTIONPVANET: Deep but Lightweight Neural Networks for Real-time Object Detection

• OBJECT DETECTIONR-FCN: Object Detection via Region-based Fully Convolutional Networks
• OBJECT DETECTIONFaster R-CNN
• OBJECT DETECTIONFast R-CNN
• OBJECT DETECTIONRich feature hierarchies for accurate object detection and semantic segmentation (aka R-CNN)
• PEDESTRIANSTen Years of Pedestrian Detection, What Have We Learned?
• NEURAL STYLEInstance Normalization: The Missing Ingredient for Fast Stylization

• HUMAN POSE ESTIMATIONStacked Hourglass Networks for Human Pose Estimation
• FACESDeepFace: Closing the Gap to Human-Level Performance in Face Verification
• TRANSLATIONCharacter-based Neural Machine Translation

• HUMAN POSE ESTIMATIONConvolutional Pose Machines
• FACESHyperFace: A Deep Multi-task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition
• FACESFace Attribute Prediction Using Off-the-Shelf CNN Features
• FACESCMS-RCNN: Contextual Multi-Scale Region-based CNN for Unconstrained Face Detection
• Conditional Image Generation with PixelCNN Decoders
• GANInfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
• GANImproved Techniques for Training GANs
• Synthesizing the preferred inputs for neurons in neural networks via deep generator networks

• ARCHITECTURESFractalNet: Ultra-Deep Neural Networks without Residuals
• PlaNet - Photo Geolocation with Convolutional Neural Networks
• OPTIMIZERSAdam: A Method for Stochastic Optimization
• GANRNNGenerating images with recurrent adversarial networks
• GANAdversarially Learned Inference

• ARCHITECTURESResnet in Resnet: Generalizing Residual Architectures
• AUTOENCODERSRank Ordered Autoencoders
• ARCHITECTURESWide Residual Networks
• ARCHITECTURESIdentity Mappings in Deep Residual Networks
• REGULARIZATIONSwapout: Learning an ensemble of deep architectures
• Multi-Scale Context Aggregation by Dilated Convolutions
• Texture Synthesis Through Convolutional Neural Networks and Spectrum Constraints
• Precomputed Real-Time Texture Synthesis with Markovian Generative Adversarial Networks

• NEURAL STYLESemantic Style Transfer and Turning Two-Bit Doodles into Fine Artwork
• Combining Markov Random Fields and Convolutional Neural Networks for Image Synthesis
• SUPERRESOLUTIONAccurate Image Super-Resolution Using Very Deep Convolutional Networks
• HUMAN POSE ESTIMATIONJoint Training of a Convolutional Network and a Graphical Model for Human Pose Estimation
• REINFORCEMENTHierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation
• COLORIZATIONLet there be Color

• NEURAL STYLEArtistic Style Transfer for Videos

• REINFORCEMENTPlaying Atari with Deep Reinforcement Learning
• GENERATIVEAttend, Infer, Repeat: Fast Scene Understanding with Generative Models
• ARCHITECTURESEFFICIENT NETWORKSSqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• ACTIVATION FUNCTIONSNoisy Activation Functions
• OBJECT DETECTIONIMAGE TO TEXTDenseCap: Fully Convolutional Localization Networks for Dense Captioning

• REGULARIZATIONDeep Networks with Stochastic Depth

• GANDeep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• GENERATIVERNNATTENTIONDRAW A Recurrent Neural Network for Image Generation
• Generating Images with Perceptual Similarity Metrics based on Deep Networks
• GENERATIVEGenerative Moment Matching Networks
• GENERATIVERNNPixel Recurrent Neural Networks
• GANUnsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

• NEURAL STYLEA Neural Algorithm for Artistic Style
• NORMALIZATIONREGULARIZATIONBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• ARCHITECTURESDeep Residual Learning for Image Recognition
• ACTIVATION FUNCTIONSFast and Accurate Deep Networks Learning By Exponential Linear Units (ELUs)
• Fractional Max-Pooling
• GANGenerative Adversarial Networks
• ARCHITECTURESInception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
• NORMALIZATIONWeight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks