Deep Learning Resources
ImageNet
Single-model on 224x224
Method | top1 | top5 | Model Size | Speed |
---|---|---|---|---|
ResNet-101 | 78.0% | 94.0% | ||
ResNet-200 | 78.3% | 94.2% | ||
Inception-v3 | ||||
Inception-v4 | ||||
Inception-ResNet-v2 | ||||
ResNet-50 | 77.8% | |||
ResNet-101 | 79.6% | 94.7% |
Single-model on 320×320 / 299×299
Method | top1 | top5 | Model Size | Speed |
---|---|---|---|---|
ResNet-101 | ||||
ResNet-200 | 79.9% | 95.2% | ||
Inception-v3 | 78.8% | 94.4% | ||
Inception-v4 | 80.0% | 95.0% | ||
Inception-ResNet-v2 | 80.1% | 95.1% | ||
ResNet-50 | ||||
ResNet-101 | 80.9% | 95.6% |
AlexNet
ImageNet Classification with Deep Convolutional Neural Networks
- nips-page: http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-
- paper: http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf
- slides: http://www.image-net.org/challenges/LSVRC/2012/supervision.pdf
- code: https://code.google.com/p/cuda-convnet/
- github: https://github.com/dnouri/cuda-convnet
- code: https://code.google.com/p/cuda-convnet2/
Network In Network
Network In Network
- intro: ICLR 2014
- arxiv: http://arxiv.org/abs/1312.4400
- gitxiv: http://gitxiv.com/posts/PA98qGuMhsijsJzgX/network-in-network-nin
- code(Caffe, official): https://gist.github.com/mavenlin/d802a5849de39225bcc6
Batch-normalized Maxout Network in Network
GoogLeNet (Inception V1)
Going Deeper with Convolutions
- arxiv: http://arxiv.org/abs/1409.4842
- github: https://github.com/google/inception
- github: https://github.com/soumith/inception.torch
Building a deeper understanding of images
VGGNet
Very Deep Convolutional Networks for Large-Scale Image Recognition
- homepage: http://www.robots.ox.ac.uk/~vgg/research/very_deep/
- arxiv: http://arxiv.org/abs/1409.1556
- slides: http://llcao.net/cu-deeplearning15/presentation/cc3580_Simonyan.pptx
- slides: http://www.robots.ox.ac.uk/~karen/pdf/ILSVRC_2014.pdf
- slides: http://deeplearning.cs.cmu.edu/slides.2015/25.simonyan.pdf
- github(official, deprecated Caffe API): https://gist.github.com/ksimonyan/211839e770f7b538e2d8
- github: https://github.com/ruimashita/caffe-train
Tensorflow VGG16 and VGG19
RepVGG: Making VGG-style ConvNets Great Again
- intro: BNRist & Tsinghua University & MEGVII Technology & Hong Kong University of Science and Technology & Aberystwyth University
- arxiv: https://arxiv.org/abs/2101.03697
- github: https://github.com/DingXiaoH/RepVGG
- github: https://github.com/megvii-model/RepVGG
Inception-V2
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- intro: ImageNet top-5 error: 4.82%
- keywords: internal covariate shift problem
- arxiv: http://arxiv.org/abs/1502.03167
- blog: https://standardfrancis.wordpress.com/2015/04/16/batch-normalization/
- notes: http://blog.csdn.net/happynear/article/details/44238541
- github: https://github.com/lim0606/caffe-googlenet-bn
ImageNet pre-trained models with batch normalization
- arxiv: https://arxiv.org/abs/1612.01452
- project page: http://www.inf-cv.uni-jena.de/Research/CNN+Models.html
- github: https://github.com/cvjena/cnn-models
Inception-V3
Inception-V3 = Inception-V2 + BN-auxiliary (fully connected layer of the auxiliary classifier is also batch-normalized, not just the convolutions)
Rethinking the Inception Architecture for Computer Vision
- intro: “21.2% top-1 and 5.6% top-5 error for single frame evaluation using a network; 3.5% top-5 error and 17.3% top-1 error With an ensemble of 4 models and multi-crop evaluation.”
- arxiv: http://arxiv.org/abs/1512.00567
- github: https://github.com/Moodstocks/inception-v3.torch
Inception in TensorFlow
- intro: demonstrate how to train the Inception v3 architecture
- github: https://github.com/tensorflow/models/tree/master/inception
Train your own image classifier with Inception in TensorFlow
- intro: Inception-v3
- blog: https://research.googleblog.com/2016/03/train-your-own-image-classifier-with.html
Notes on the TensorFlow Implementation of Inception v3
Training an InceptionV3-based image classifier with your own dataset
Inception-BN full for Caffe: Inception-BN ImageNet (21K classes) model for Caffe
ResNet
Deep Residual Learning for Image Recognition
- intro: CVPR 2016 Best Paper Award
- arxiv: http://arxiv.org/abs/1512.03385
- slides: http://research.microsoft.com/en-us/um/people/kahe/ilsvrc15/ilsvrc2015_deep_residual_learning_kaiminghe.pdf
- gitxiv: http://gitxiv.com/posts/LgPRdTY3cwPBiMKbm/deep-residual-learning-for-image-recognition
- github: https://github.com/KaimingHe/deep-residual-networks
- github: https://github.com/ry/tensorflow-resnet
Third-party re-implementations
https://github.com/KaimingHe/deep-residual-networks#third-party-re-implementations
Training and investigating Residual Nets
- intro: Facebook AI Research
- blog: http://torch.ch/blog/2016/02/04/resnets.html
- github: https://github.com/facebook/fb.resnet.torch
resnet.torch: an updated version of fb.resnet.torch with many changes.
Highway Networks and Deep Residual Networks
Interpretating Deep Residual Learning Blocks as Locally Recurrent Connections
Lab41 Reading Group: Deep Residual Learning for Image Recognition
50-layer ResNet, trained on ImageNet, classifying webcam
- homepage: https://ml4a.github.io/demos/keras.js/
Reproduced ResNet on CIFAR-10 and CIFAR-100 dataset.
ResNet-V2
Identity Mappings in Deep Residual Networks
- intro: ECCV 2016. ResNet-v2
- arxiv: http://arxiv.org/abs/1603.05027
- github: https://github.com/KaimingHe/resnet-1k-layers
- github: https://github.com/tornadomeet/ResNet
Deep Residual Networks for Image Classification with Python + NumPy
Inception-V4 / Inception-ResNet-V2
Inception-V4, Inception-Resnet And The Impact Of Residual Connections On Learning
- intro: Workshop track - ICLR 2016. 3.08 % top-5 error on ImageNet CLS
- intro: “achieve 3.08% top-5 error on the test set of the ImageNet classification (CLS) challenge”
- arxiv: http://arxiv.org/abs/1602.07261
- github(Keras): https://github.com/kentsommer/keras-inceptionV4
The inception-resnet-v2 models trained from scratch via torch
Inception v4 in Keras
- intro: Inception-v4, Inception - Resnet-v1 and v2
- github: https://github.com/titu1994/Inception-v4
ResNeXt
Aggregated Residual Transformations for Deep Neural Networks
- intro: CVPR 2017. UC San Diego & Facebook AI Research
- arxiv: https://arxiv.org/abs/1611.05431
- github(Torch): https://github.com/facebookresearch/ResNeXt
- github: https://github.com/dmlc/mxnet/blob/master/example/image-classification/symbol/resnext.py
- dataset: http://data.dmlc.ml/models/imagenet/resnext/
- reddit: https://www.reddit.com/r/MachineLearning/comments/5haml9/p_implementation_of_aggregated_residual/
ResNeSt
ResNeSt: Split-Attention Networks
- intro: Amazon & University of California
- arxiv: https://arxiv.org/abs/2004.08955
- github: https://github.com/zhanghang1989/ResNeSt
Residual Networks Variants
Resnet in Resnet: Generalizing Residual Architectures
- paper: http://beta.openreview.net/forum?id=lx9l4r36gU2OVPy8Cv9g
- arxiv: http://arxiv.org/abs/1603.08029
Residual Networks are Exponential Ensembles of Relatively Shallow Networks
Wide Residual Networks
- intro: BMVC 2016
- arxiv: http://arxiv.org/abs/1605.07146
- github: https://github.com/szagoruyko/wide-residual-networks
- github: https://github.com/asmith26/wide_resnets_keras
- github: https://github.com/ritchieng/wideresnet-tensorlayer
- github: https://github.com/xternalz/WideResNet-pytorch
- github(Torch): https://github.com/meliketoy/wide-residual-network
Residual Networks of Residual Networks: Multilevel Residual Networks
Multi-Residual Networks
Deep Pyramidal Residual Networks
- intro: PyramidNet
- arxiv: https://arxiv.org/abs/1610.02915
- github: https://github.com/jhkim89/PyramidNet
Learning Identity Mappings with Residual Gates
Wider or Deeper: Revisiting the ResNet Model for Visual Recognition
- intro: image classification, semantic image segmentation
- arxiv: https://arxiv.org/abs/1611.10080
- github: https://github.com/itijyou/ademxapp
Deep Pyramidal Residual Networks with Separated Stochastic Depth
Spatially Adaptive Computation Time for Residual Networks
- intro: Higher School of Economics & Google & CMU
- arxiv: https://arxiv.org/abs/1612.02297
ShaResNet: reducing residual network parameter number by sharing weights
Sharing Residual Units Through Collective Tensor Factorization in Deep Neural Networks
- intro: Collective Residual Networks
- arxiv: https://arxiv.org/abs/1703.02180
- github(MXNet): https://github.com/cypw/CRU-Net
Residual Attention Network for Image Classification
- intro: CVPR 2017 Spotlight. SenseTime Group Limited & Tsinghua University & The Chinese University of Hong Kong
- intro: ImageNet (4.8% single model and single crop, top-5 error)
- arxiv: https://arxiv.org/abs/1704.06904
- github(Caffe): https://github.com/buptwangfei/residual-attention-network
Dilated Residual Networks
- intro: CVPR 2017. Princeton University & Intel Labs
- keywords: Dilated Residual Networks (DRN)
- project page: http://vladlen.info/publications/dilated-residual-networks/
- arxiv: https://arxiv.org/abs/1705.09914
- paper: http://vladlen.info/papers/DRN.pdf
Dynamic Steerable Blocks in Deep Residual Networks
- intro: University of Amsterdam & ESAT-PSI
- arxiv: https://arxiv.org/abs/1706.00598
Learning Deep ResNet Blocks Sequentially using Boosting Theory
- intro: Microsoft Research & Princeton University
- arxiv: https://arxiv.org/abs/1706.04964
Learning Strict Identity Mappings in Deep Residual Networks
- keywords: epsilon-ResNet
- arxiv: https://arxiv.org/abs/1804.01661
Spiking Deep Residual Network
https://arxiv.org/abs/1805.01352
Norm-Preservation: Why Residual Networks Can Become Extremely Deep?
- intro: University of Central Florida
- arxiv: https://arxiv.org/abs/1805.07477
MEAL V2: Boosting Vanilla ResNet-50 to 80%+ Top-1 Accuracy on ImageNet without Tricks
- intro: Carnegie Mellon University
- arxiv: https://arxiv.org/abs/2009.08453
- github: https://github.com/szq0214/MEAL-V2
DenseNet
Densely Connected Convolutional Networks
- intro: CVPR 2017 best paper. Cornell University & Tsinghua University. DenseNet
- arxiv: http://arxiv.org/abs/1608.06993
- github: https://github.com/liuzhuang13/DenseNet
- github(Lasagne): https://github.com/Lasagne/Recipes/tree/master/papers/densenet
- github(Keras): https://github.com/tdeboissiere/DeepLearningImplementations/tree/master/DenseNet
- github(Caffe): https://github.com/liuzhuang13/DenseNetCaffe
- github(Tensorflow): https://github.com/YixuanLi/densenet-tensorflow
- github(Keras): https://github.com/titu1994/DenseNet
- github(PyTorch): https://github.com/bamos/densenet.pytorch
- github(PyTorch): https://github.com/andreasveit/densenet-pytorch
- github(Tensorflow): https://github.com/ikhlestov/vision_networks
Memory-Efficient Implementation of DenseNets
- intro: Cornell University & Fudan University & Facebook AI Research
- arxiv: https://arxiv.org/abs/1707.06990
- github: https://github.com/liuzhuang13/DenseNet/tree/master/models
- github: https://github.com/gpleiss/efficient_densenet_pytorch
- github: https://github.com/taineleau/efficient_densenet_mxnet
- github: https://github.com/Tongcheng/DN_CaffeScript
DenseNet 2.0
CondenseNet: An Efficient DenseNet using Learned Group Convolutions
Multimodal Densenet
https://arxiv.org/abs/1811.07407
Xception
Deep Learning with Separable Convolutions
Xception: Deep Learning with Depthwise Separable Convolutions
- intro: CVPR 2017. Extreme Inception
- arxiv: https://arxiv.org/abs/1610.02357
- code: https://keras.io/applications/#xception
- github(Keras): https://github.com/fchollet/deep-learning-models/blob/master/xception.py
- github: https://gist.github.com/culurciello/554c8e56d3bbaf7c66bf66c6089dc221
- github: https://github.com/kwotsin/Tensorflow-Xception
- github: https://github.com//bruinxiong/xception.mxnet
- notes: http://www.shortscience.org/paper?bibtexKey=journals%2Fcorr%2F1610.02357
Towards a New Interpretation of Separable Convolutions
MobileNets
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
- intro: Google
- arxiv: https://arxiv.org/abs/1704.04861
- github: https://github.com/rcmalli/keras-mobilenet
- github: https://github.com/marvis/pytorch-mobilenet
- github(Tensorflow): https://github.com/Zehaos/MobileNet
- github: https://github.com/shicai/MobileNet-Caffe
- github: https://github.com/hollance/MobileNet-CoreML
- github: https://github.com/KeyKy/mobilenet-mxnet
MobileNets: Open-Source Models for Efficient On-Device Vision
- blog: https://research.googleblog.com/2017/06/mobilenets-open-source-models-for.html
- github: https://github.com/tensorflow/models/blob/master/slim/nets/mobilenet_v1.md
Google’s MobileNets on the iPhone
- blog: http://machinethink.net/blog/googles-mobile-net-architecture-on-iphone/
- github: https://github.com/hollance/MobileNet-CoreML
Depth_conv-for-mobileNet
https://github.com//LamHoCN/Depth_conv-for-mobileNet
The Enhanced Hybrid MobileNet
https://arxiv.org/abs/1712.04698
FD-MobileNet: Improved MobileNet with a Fast Downsampling Strategy
https://arxiv.org/abs/1802.03750
A Quantization-Friendly Separable Convolution for MobileNets
- intro: THE 1ST WORKSHOP ON ENERGY EFFICIENT MACHINE LEARNING AND COGNITIVE COMPUTING FOR EMBEDDED APPLICATIONS (EMC2)
- arxiv: https://arxiv.org/abs/1803.08607
MobileNetV2
Inverted Residuals and Linear Bottlenecks: Mobile Networks forClassification, Detection and Segmentation
- intro: Google
- keywords: MobileNetV2, SSDLite, DeepLabv3
- arxiv: https://arxiv.org/abs/1801.04381
- github: https://github.com/tensorflow/models/tree/master/research/slim/nets/mobilenet
- github: https://github.com/liangfu/mxnet-mobilenet-v2
- blog: https://research.googleblog.com/2018/04/mobilenetv2-next-generation-of-on.html
PydMobileNet: Improved Version of MobileNets with Pyramid Depthwise Separable Convolution
https://arxiv.org/abs/1811.07083
Rethinking Depthwise Separable Convolutions: How Intra-Kernel Correlations Lead to Improved MobileNets
- intro: CVPR 2020
- arxiv: https://arxiv.org/abs/2003.13549
- github: https://github.com/zeiss-microscopy/BSConv
Rethinking Bottleneck Structure for Efficient Mobile Network Design
- intro: ECCV 2020
- intro: National University of Singapore & Yitu Technology
- arxiv: https://arxiv.org/abs/2007.02269
- github: https://github.com/zhoudaquan/rethinking_bottleneck_design
Mobile-Former: Bridging MobileNet and Transformer
- intro: Microsoft & University of Science and Technology of China
- arxiv: https://arxiv.org/abs/2108.05895
ShuffleNet
ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices
- intro: Megvii Inc (Face++)
- arxiv: https://arxiv.org/abs/1707.01083
ShuffleNet V2
ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
- intro: ECCV 2018. Megvii Inc (Face++) & Tsinghua University
- arxiv: [https://arxiv.org/abs/1807.11164](https://arxiv.org/abs/1807.11164
SENet
Squeeze-and-Excitation Networks
- intro: CVPR 2018
- intro: ILSVRC 2017 image classification winner. Momenta & University of Oxford
- arxiv: https://arxiv.org/abs/1709.01507
- github(official, Caffe): https://github.com/hujie-frank/SENet
- github: https://github.com/bruinxiong/SENet.mxnet
Competitive Inner-Imaging Squeeze and Excitation for Residual Network
GENet
Gather-Excite: Exploiting Feature Context in Convolutional Neural Networks
- intro: NIPS 2018
- github: https://github.com/hujie-frank/GENet
A ConvNet for the 2020s
- intro: Facebook AI Research (FAIR) & UC Berkeley
- arxiv: https://arxiv.org/abs/2201.03545
- github: https://github.com/facebookresearch/ConvNeXt
ImageNet Projects
Training an Object Classifier in Torch-7 on multiple GPUs over ImageNet
- intro: an imagenet example in torch
- github: https://github.com/soumith/imagenet-multiGPU.torch
Pre-training
Exploring the Limits of Weakly Supervised Pretraining
- intro: report the highest ImageNet-1k single-crop, top-1 accuracy to date: 85.4% (97.6% top-5)
- paper: https://research.fb.com/publications/exploring-the-limits-of-weakly-supervised-pretraining/
Rethinking ImageNet Pre-training
- intro: Facebook AI Research
- arxiv: https://arxiv.org/abs/1811.08883
Revisiting Pre-training: An Efficient Training Method for Image Classification
https://arxiv.org/abs/1811.09347
Rethinking Pre-training and Self-training
- intro: NeurIPS 2020
- intro: Google Research, Brain Team
- arxiv: https://arxiv.org/abs/2006.06882
- github: https://github.com/tensorflow/tpu/tree/master/models/official/detection/projects/self_training
Exploring the Limits of Large Scale Pre-training
- intro: Google Research
- arxiv: https://arxiv.org/abs/2110.02095
Transformers
Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet
- intro: National University of Singapore & YITU Technology
- arxiv: https://arxiv.org/abs/2101.11986
- github: https://github.com/yitu-opensource/T2T-ViT
Incorporating Convolution Designs into Visual Transformers
- intro: SenseTime Research & Nanyang Technological University
- arxiv: https://arxiv.org/abs/2103.11816
DeepViT: Towards Deeper Vision Transformer
- intro: National University of Singapore & ByteDance US AI Lab
- arxiv: https://arxiv.org/abs/2103.11886
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
- intro: ICCV 2021 best paper
- intro: Microsoft Research Asia
- arxiv: https://arxiv.org/abs/2103.14030
- github: https://github.com/microsoft/Swin-Transformer
Rethinking the Design Principles of Robust Vision Transformer
- arxiv: https://arxiv.org/abs/2105.07926
- github: https://github.com/vtddggg/Robust-Vision-Transformer
Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers
- intro: Google Research & DeepMind
- arxiv: https://arxiv.org/abs/2109.10686
How Do Vision Transformers Work?
- intro: ICLR 2022 Spotlight
- intro: Yonsei University & NAVER AI Lab
- arxiv: https://arxiv.org/abs/2202.06709
- github: https://github.com/xxxnell/how-do-vits-work
MulT: An End-to-End Multitask Learning Transformer
- intro: CVPR 2022
- project page: https://ivrl.github.io/MulT/
- arxiv: https://arxiv.org/abs/2205.08303
EfficientFormer: Vision Transformers at MobileNet Speed
- intro: Snap Inc. & Northeastern University
- arxiv: https://arxiv.org/abs/2206.01191
- github: https://github.com/snap-research/EfficientFormer
SimA: Simple Softmax-free Attention for Vision Transformers
- intro: University of Maryland & University of California
- arxiv: https://arxiv.org/abs/2206.08898
- gihtub: https://github.com/UCDvision/sima
Semi-Supervised Learning
Semi-Supervised Learning with Graphs
- intro: Label Propagation
- paper: http://pages.cs.wisc.edu/~jerryzhu/pub/thesis.pdf
- blog(“标签传播算法(Label Propagation)及Python实现”): http://blog.csdn.net/zouxy09/article/details/49105265
Semi-Supervised Learning with Ladder Networks
- arxiv: http://arxiv.org/abs/1507.02672
- github: https://github.com/CuriousAI/ladder
- github: https://github.com/rinuboney/ladder
Semi-supervised Feature Transfer: The Practical Benefit of Deep Learning Today?
Temporal Ensembling for Semi-Supervised Learning
- intro: ICLR 2017
- arxiv: https://arxiv.org/abs/1610.02242
- github: https://github.com/smlaine2/tempens
Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data
- intro: ICLR 2017 best paper award
- arxiv: https://arxiv.org/abs/1610.05755
- github: https://github.com/tensorflow/models/tree/8505222ea1f26692df05e65e35824c6c71929bb5/privacy
Infinite Variational Autoencoder for Semi-Supervised Learning
Multi-label Learning
CNN: Single-label to Multi-label
Deep Learning for Multi-label Classification
- arxiv: http://arxiv.org/abs/1502.05988
- github: http://meka.sourceforge.net
Predicting Unseen Labels using Label Hierarchies in Large-Scale Multi-label Learning
- intro: ECML 2015
- paper: https://www.kdsl.tu-darmstadt.de/fileadmin/user_upload/Group_KDSL/PUnL_ECML2015_camera_ready.pdf
Learning with a Wasserstein Loss
- project page: http://cbcl.mit.edu/wasserstein/
- arxiv: http://arxiv.org/abs/1506.05439
- code: http://cbcl.mit.edu/wasserstein/yfcc100m_labels.tar.gz
- MIT news: http://news.mit.edu/2015/more-flexible-machine-learning-1001
From Softmax to Sparsemax: A Sparse Model of Attention and Multi-Label Classification
- intro: ICML 2016
- arxiv: http://arxiv.org/abs/1602.02068
- github: https://github.com/gokceneraslan/SparseMax.torch
- github: https://github.com/Unbabel/sparsemax
CNN-RNN: A Unified Framework for Multi-label Image Classification
Improving Multi-label Learning with Missing Labels by Structured Semantic Correlations
Extreme Multi-label Loss Functions for Recommendation, Tagging, Ranking & Other Missing Label Applications
- intro: Indian Institute of Technology Delhi & MSR
- paper: https://manikvarma.github.io/pubs/jain16.pdf
Multi-Label Image Classification with Regional Latent Semantic Dependencies
- intro: Regional Latent Semantic Dependencies model (RLSD), RNN, RPN
- arxiv: https://arxiv.org/abs/1612.01082
Privileged Multi-label Learning
- intro: Peking University & University of Technology Sydney & University of Sydney
- arxiv: https://arxiv.org/abs/1701.07194
Multi-task Learning
Multitask Learning / Domain Adaptation
multi-task learning
- discussion: https://github.com/memect/hao/issues/93
Learning and Transferring Multi-task Deep Representation for Face Alignment
Multi-task learning of facial landmarks and expression
Multi-Task Deep Visual-Semantic Embedding for Video Thumbnail Selection
- intro: CVPR 2015
- paper: http://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Liu_Multi-Task_Deep_Visual-Semantic_2015_CVPR_paper.pdf
Learning Multiple Tasks with Deep Relationship Networks
Learning deep representation of multityped objects and tasks
Cross-stitch Networks for Multi-task Learning
Multi-Task Learning in Tensorflow (Part 1)
Deep Multi-Task Learning with Shared Memory
- intro: EMNLP 2016
- arxiv: http://arxiv.org/abs/1609.07222
Learning to Push by Grasping: Using multiple tasks for effective learning
Identifying beneficial task relations for multi-task learning in deep neural networks
- intro: EACL 2017
- arxiv: https://arxiv.org/abs/1702.08303
- github: https://github.com/jbingel/eacl2017_mtl
Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics
- intro: University of Cambridge
- arxiv: https://arxiv.org/abs/1705.07115
One Model To Learn Them All
- intro: Google Brain & University of Toronto
- arxiv: https://arxiv.org/abs/1706.05137
- github: https://github.com/tensorflow/tensor2tensor
MultiModel: Multi-Task Machine Learning Across Domains
https://research.googleblog.com/2017/06/multimodel-multi-task-machine-learning.html
An Overview of Multi-Task Learning in Deep Neural Networks
- intro: Aylien Ltd
- arxiv: https://arxiv.org/abs/1706.05098
PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning
End-to-End Multi-Task Learning with Attention
- intro: Imperial College London
- arxiv: https://arxiv.org/abs/1803.10704
Cross-connected Networks for Multi-task Learning of Detection and Segmentation
https://arxiv.org/abs/1805.05569
Auxiliary Tasks in Multi-task Learning
https://arxiv.org/abs/1805.06334
K For The Price Of 1: Parameter Efficient Multi-task And Transfer Learning
- intro: The University of Chicago & Google
- arxiv: https://arxiv.org/abs/1810.10703
Which Tasks Should Be Learned Together in Multi-task Learning?
- intro: ICML 2020
- intro: Stanford
- project page: http://taskgrouping.stanford.edu/
- arxiv: https://arxiv.org/abs/1905.07553
OmniNet: A unified architecture for multi-modal multi-task learning
Deep Elastic Networks with Model Selection for Multi-Task Learning
- intro: ICCV 2019
- arxiv: https://arxiv.org/abs/1909.04860
AdaShare: Learning What To Share For Efficient Deep Multi-Task Learning
- intro: Boston University & IBM Research & MIT-IBM Watson AI Lab
- arxiv: https://arxiv.org/abs/1911.12423
Multi-Task Learning for Dense Prediction Tasks: A Survey
- intro: T-PAMI
- arxiv: https://arxiv.org/abs/2004.13379
- github: https://github.com/SimonVandenhende/Multi-Task-Learning-PyTorch
MTI-Net: Multi-Scale Task Interaction Networks for Multi-Task Learning
- intro: ECCV 2020 spotlight
- keywords: MTI-Net
- arxiv: https://arxiv.org/abs/2001.06902
- github: https://github.com/SimonVandenhende/Multi-Task-Learning-PyTorch
Exploring Relational Context for Multi-Task Dense Prediction
- intro: ETH Zurich
- arxiv: https://arxiv.org/abs/2104.13874
Cross-task Attention Mechanism for Dense Multi-task Learning
- intro: Inria, France & Valeo.ai, France
- arxiv: https://arxiv.org/abs/2206.08927
- github: https://github.com/cv-rits/DenseMTL
Multi-modal Learning
Multimodal Deep Learning
Multimodal Convolutional Neural Networks for Matching Image and Sentence
- homepage: http://mcnn.noahlab.com.hk/project.html
- paper: http://mcnn.noahlab.com.hk/ICCV2015.pdf
- arxiv: http://arxiv.org/abs/1504.06063
A C++ library for Multimodal Deep Learning
Multimodal Learning for Image Captioning and Visual Question Answering
Multi modal retrieval and generation with deep distributed models
- slides: http://www.slideshare.net/roelofp/multi-modal-retrieval-and-generation-with-deep-distributed-models
- slides: http://pan.baidu.com/s/1kUSjn4z
Learning Aligned Cross-Modal Representations from Weakly Aligned Data
- homepage: http://projects.csail.mit.edu/cmplaces/index.html
- paper: http://projects.csail.mit.edu/cmplaces/content/paper.pdf
Variational methods for Conditional Multimodal Deep Learning
Training and Evaluating Multimodal Word Embeddings with Large-scale Web Annotated Images
- intro: NIPS 2016. University of California & Pinterest
- project page: http://www.stat.ucla.edu/~junhua.mao/multimodal_embedding.html
- arxiv: https://arxiv.org/abs/1611.08321
Deep Multi-Modal Image Correspondence Learning
Multimodal Deep Learning (D4L4 Deep Learning for Speech and Language UPC 2017)
Multimodal Learning with Transformers: A Survey
- intro: University of Oxford & University of Surrey
- arxiv: https://arxiv.org/abs/2206.06488
Debugging Deep Learning
Some tips for debugging deep learning
Introduction to debugging neural networks
- blog: http://russellsstewart.com/notes/0.html
- reddit: https://www.reddit.com/r/MachineLearning/comments/4du7gv/introduction_to_debugging_neural_networks
How to Visualize, Monitor and Debug Neural Network Learning
Learning from learning curves
- intro: Kaggle
- blog: https://medium.com/@dsouza.amanda/learning-from-learning-curves-1a82c6f98f49#.o5synrvvl
Understanding CNN
Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
- intro: NIPS 2016
- paper: http://www.cs.toronto.edu/~wenjie/papers/nips16/top.pdf
Deep Learning Networks
PCANet: A Simple Deep Learning Baseline for Image Classification?
- arixv: http://arxiv.org/abs/1404.3606
- code(Matlab): http://mx.nthu.edu.tw/~tsunghan/download/PCANet_demo_pyramid.rar
- mirror: http://pan.baidu.com/s/1mg24b3a
- github(C++): https://github.com/Ldpe2G/PCANet
- github(Python): https://github.com/IshitaTakeshi/PCANet
Convolutional Kernel Networks
- intro: NIPS 2014
- arxiv: http://arxiv.org/abs/1406.3332
Deeply-supervised Nets
- intro: DSN
- arxiv: http://arxiv.org/abs/1409.5185
- homepage: http://vcl.ucsd.edu/~sxie/2014/09/12/dsn-project/
- github: https://github.com/s9xie/DSN
- notes: http://zhangliliang.com/2014/11/02/paper-note-dsn/
FitNets: Hints for Thin Deep Nets
Striving for Simplicity: The All Convolutional Net
- intro: ICLR-2015 workshop
- arxiv: http://arxiv.org/abs/1412.6806
How these researchers tried something unconventional to come out with a smaller yet better Image Recognition.
- intro: All Convolutional Network: (https://arxiv.org/abs/1412.6806#) implementation in Keras
- blog: https://medium.com/@matelabs_ai/how-these-researchers-tried-something-unconventional-to-came-out-with-a-smaller-yet-better-image-544327f30e72#.pfdbvdmuh
- blog: https://github.com/MateLabs/All-Conv-Keras
Pointer Networks
- arxiv: https://arxiv.org/abs/1506.03134
- github: https://github.com/vshallc/PtrNets
- github(TensorFlow): https://github.com/ikostrikov/TensorFlow-Pointer-Networks
- github(TensorFlow): https://github.com/devsisters/pointer-network-tensorflow
- notes: https://github.com/dennybritz/deeplearning-papernotes/blob/master/notes/pointer-networks.md
Pointer Networks in TensorFlow (with sample code)
- blog: https://medium.com/@devnag/pointer-networks-in-tensorflow-with-sample-code-14645063f264#.sxipqfj30
- github: https://github.com/devnag/tensorflow-pointer-networks
Rectified Factor Networks
- arxiv: http://arxiv.org/abs/1502.06464
- github: https://github.com/untom/librfn
Correlational Neural Networks
Diversity Networks
Competitive Multi-scale Convolution
A Unified Approach for Learning the Parameters of Sum-Product Networks (SPN)
- intro: “The Sum-Product Network (SPN) is a new type of machine learning model with fast exact probabilistic inference over many layers.”
- arxiv: http://arxiv.org/abs/1601.00318
- homepage: http://spn.cs.washington.edu/index.shtml
- code: http://spn.cs.washington.edu/code.shtml
Awesome Sum-Product Networks
Recombinator Networks: Learning Coarse-to-Fine Feature Aggregation
- intro: CVPR 2016
- arxiv: http://arxiv.org/abs/1511.07356
- paper: http://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Honari_Recombinator_Networks_Learning_CVPR_2016_paper.pdf
- github: https://github.com/SinaHonari/RCN
Dynamic Capacity Networks
- intro: ICML 2016
- arxiv: http://arxiv.org/abs/1511.07838
- github(Tensorflow): https://github.com/beopst/dcn.tf
- review: http://www.erogol.com/1314-2/
Bitwise Neural Networks
- paper: http://paris.cs.illinois.edu/pubs/minje-icmlw2015.pdf
- demo: http://minjekim.com/demo_bnn.html
Learning Discriminative Features via Label Consistent Neural Network
A Theory of Generative ConvNet
- project page: http://www.stat.ucla.edu/~ywu/GenerativeConvNet/main.html
- arxiv: http://arxiv.org/abs/1602.03264
- code: http://www.stat.ucla.edu/~ywu/GenerativeConvNet/doc/code.zip
How to Train Deep Variational Autoencoders and Probabilistic Ladder Networks
Group Equivariant Convolutional Networks (G-CNNs)
Deep Spiking Networks
Low-rank passthrough neural networks
Single Image 3D Interpreter Network
- intro: ECCV 2016 (oral)
- arxiv: https://arxiv.org/abs/1604.08685
Deeply-Fused Nets
SNN: Stacked Neural Networks
Universal Correspondence Network
- intro: NIPS 2016 full oral presentation. Stanford University & NEC Laboratories America
- project page: http://cvgl.stanford.edu/projects/ucn/
- arxiv: https://arxiv.org/abs/1606.03558
Progressive Neural Networks
- intro: Google DeepMind
- arxiv: https://arxiv.org/abs/1606.04671
- github: https://github.com/synpon/prog_nn
- github: https://github.com/yao62995/A3C
Holistic SparseCNN: Forging the Trident of Accuracy, Speed, and Size
Mollifying Networks
- author: Caglar Gulcehre, Marcin Moczulski, Francesco Visin, Yoshua Bengio
- arxiv: http://arxiv.org/abs/1608.04980
Domain Separation Networks
- intro: NIPS 2016
- intro: Google Brain & Imperial College London & Google Research
- arxiv: https://arxiv.org/abs/1608.06019
- github: https://github.com/tensorflow/models/tree/master/domain_adaptation
Local Binary Convolutional Neural Networks
CliqueCNN: Deep Unsupervised Exemplar Learning
- intro: NIPS 2016
- arxiv: http://arxiv.org/abs/1608.08792
- github: https://github.com/asanakoy/cliquecnn
Convexified Convolutional Neural Networks
Multi-scale brain networks
https://arxiv.org/abs/1711.11473
Input Convex Neural Networks
- arxiv: http://arxiv.org/abs/1609.07152
- github: https://github.com/locuslab/icnn
HyperNetworks
- arxiv: https://arxiv.org/abs/1609.09106
- blog: http://blog.otoro.net/2016/09/28/hyper-networks/
- github: https://github.com/hardmaru/supercell/blob/master/assets/MNIST_Static_HyperNetwork_Example.ipynb
HyperLSTM
X-CNN: Cross-modal Convolutional Neural Networks for Sparse Datasets
Tensor Switching Networks
- intro: NIPS 2016
- arixiv: https://arxiv.org/abs/1610.10087
- github: https://github.com/coxlab/tsnet
BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks
- intro: Harvard University
- paper: http://www.eecs.harvard.edu/~htk/publication/2016-icpr-teerapittayanon-mcdanel-kung.pdf
- github: https://github.com/kunglab/branchynet
Spectral Convolution Networks
DelugeNets: Deep Networks with Massive and Flexible Cross-layer Information Inflows
PolyNet: A Pursuit of Structural Diversity in Very Deep Networks
- arxiv: https://arxiv.org/abs/1611.05725
- poster: http://mmlab.ie.cuhk.edu.hk/projects/cu_deeplink/polynet_poster.pdf
Weakly Supervised Cascaded Convolutional Networks
DeepSetNet: Predicting Sets with Deep Neural Networks
- intro: multi-class image classification and pedestrian detection
- arxiv: https://arxiv.org/abs/1611.08998
Steerable CNNs
- intro: University of Amsterdam
- arxiv: https://arxiv.org/abs/1612.08498
Feedback Networks
- project page: http://feedbacknet.stanford.edu/
- arxiv: https://arxiv.org/abs/1612.09508
- youtube: https://youtu.be/MY5Uhv38Ttg
Oriented Response Networks
OptNet: Differentiable Optimization as a Layer in Neural Networks
A fast and differentiable QP solver for PyTorch
- github: https://github.com/locuslab/qpth
Meta Networks
https://arxiv.org/abs/1703.00837
Deformable Convolutional Networks
- intro: ICCV 2017 oral. Microsoft Research Asia
- keywords: deformable convolution, deformable RoI pooling
- arxiv: https://arxiv.org/abs/1703.06211
- sliedes: http://www.jifengdai.org/slides/Deformable_Convolutional_Networks_Oral.pdf
- github(official): https://github.com/msracver/Deformable-ConvNets
- github: https://github.com/felixlaumon/deform-conv
- github: https://github.com/oeway/pytorch-deform-conv
Deformable ConvNets v2: More Deformable, Better Results**
- intro: University of Science and Technology of China & Microsoft Research Asia
- keywords: DCNv2
- arxiv: https://arxiv.org/abs/1811.11168
- github: https://github.com/msracver/Deformable-ConvNets/tree/master/DCNv2_op
Second-order Convolutional Neural Networks
https://arxiv.org/abs/1703.06817
Gabor Convolutional Networks
https://arxiv.org/abs/1705.01450
Deep Rotation Equivariant Network
https://arxiv.org/abs/1705.08623
Dense Transformer Networks
- intro: Washington State University & University of California, Davis
- arxiv: https://arxiv.org/abs/1705.08881
- github: https://github.com/divelab/dtn
Deep Complex Networks
- intro: [Université de Montréal & INRS-EMT & Microsoft Maluuba
- arxiv: https://arxiv.org/abs/1705.09792
- github: https://github.com/ChihebTrabelsi/deep_complex_networks
Deep Quaternion Networks
- intro: University of Louisiana
- arxiv: https://arxiv.org/abs/1712.04604
DiracNets: Training Very Deep Neural Networks Without Skip-Connections
- intro: Université Paris-Est
- arxiv: https://arxiv.org/abs/1706.00388
- github: https://github.com/szagoruyko/diracnets
Dual Path Networks
- intro: National University of Singapore
- arxiv: https://arxiv.org/abs/1707.01629
- github(MXNet): https://github.com/cypw/DPNs
Primal-Dual Group Convolutions for Deep Neural Networks
Interleaved Group Convolutions for Deep Neural Networks
- intro: ICCV 2017
- keywords: interleaved group convolutional neural networks (IGCNets), IGCV1
- arxiv: https://arxiv.org/abs/1707.02725
- gihtub: https://github.com/hellozting/InterleavedGroupConvolutions
IGCV2: Interleaved Structured Sparse Convolutional Neural Networks
- intro: CVPR 2018
- arxiv: https://arxiv.org/abs/1804.06202
IGCV3: Interleaved Low-Rank Group Convolutions for Efficient Deep Neural Networks
- intro: University of Scinence and Technology of China & Microsoft Reserach Asia
- arxiv: https://arxiv.org/abs/1806.00178
- github(official): https://github.com/homles11/IGCV3
Sensor Transformation Attention Networks
https://arxiv.org/abs/1708.01015
Sparsity Invariant CNNs
https://arxiv.org/abs/1708.06500
SPARCNN: SPAtially Related Convolutional Neural Networks
https://arxiv.org/abs/1708.07522
BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks
https://arxiv.org/abs/1709.01686
Polar Transformer Networks
https://arxiv.org/abs/1709.01889
Tensor Product Generation Networks
https://arxiv.org/abs/1709.09118
Deep Competitive Pathway Networks
- intro: ACML 2017
- arxiv: https://arxiv.org/abs/1709.10282
- github: https://github.com/JiaRenChang/CoPaNet
Context Embedding Networks
https://arxiv.org/abs/1710.01691
Generalization in Deep Learning
- intro: MIT & University of Montreal
- arxiv: https://arxiv.org/abs/1710.05468
Understanding Deep Learning Generalization by Maximum Entropy
- intro: University of Science and Technology of China & Beijing Jiaotong University & Chinese Academy of Sciences
- arxiv: https://arxiv.org/abs/1711.07758
Do Convolutional Neural Networks Learn Class Hierarchy?
- intro: Bosch Research North America & Michigan State University
- arxiv: https://arxiv.org/abs/1710.06501
- video demo: https://vimeo.com/228263798
Deep Hyperspherical Learning
- intro: NIPS 2017
- arxiv: https://arxiv.org/abs/1711.03189
Beyond Sparsity: Tree Regularization of Deep Models for Interpretability
- intro: AAAI 2018
- arxiv: https://arxiv.org/abs/1711.06178
Neural Motifs: Scene Graph Parsing with Global Context
- keywords: Stacked Motif Networks
- arxiv: https://arxiv.org/abs/1711.06640
Priming Neural Networks
https://arxiv.org/abs/1711.05918
Three Factors Influencing Minima in SGD
https://arxiv.org/abs/1711.04623
BPGrad: Towards Global Optimality in Deep Learning via Branch and Pruning
https://arxiv.org/abs/1711.06959
BlockDrop: Dynamic Inference Paths in Residual Networks
- intro: UMD & UT Austin & IBM Research & Fusemachines Inc.
- arxiv: https://arxiv.org/abs/1711.08393
Wasserstein Introspective Neural Networks
https://arxiv.org/abs/1711.08875
SkipNet: Learning Dynamic Routing in Convolutional Networks
https://arxiv.org/abs/1711.09485
Do Convolutional Neural Networks act as Compositional Nearest Neighbors?
- intro: CMU & West Virginia University
- arxiv: https://arxiv.org/abs/1711.10683
ConvNets and ImageNet Beyond Accuracy: Explanations, Bias Detection, Adversarial Examples and Model Criticism
- intro: Facebook AI Research
- arxiv: https://arxiv.org/abs/1711.11443
Broadcasting Convolutional Network
https://arxiv.org/abs/1712.02517
Point-wise Convolutional Neural Network
- intro: Singapore University of Technology and Design
- arxiv: https://arxiv.org/abs/1712.05245
ScreenerNet: Learning Curriculum for Neural Networks
- intro: Intel Corporation & Allen Institute for Artificial Intelligence
- keywords: curricular learning, deep learning, deep q-learning
- arxiv: https://arxiv.org/abs/1801.00904
Sparsely Connected Convolutional Networks
https://arxiv.org/abs/1801.05895
Spherical CNNs
- intro: ICLR 2018 best paper award. University of Amsterdam & EPFL
- arxiv: https://arxiv.org/abs/1801.10130
- github(official, PyTorch): https://github.com/jonas-koehler/s2cnn
Going Deeper in Spiking Neural Networks: VGG and Residual Architectures
- intro: Purdue University & Oculus Research & Facebook Research
- arxiv: https://arxiv.org/abs/1802.02627
Rotate your Networks: Better Weight Consolidation and Less Catastrophic Forgetting
https://arxiv.org/abs/1802.02950
Convolutional Neural Networks with Alternately Updated Clique
- intro: CVPR 2018
- arxiv: https://arxiv.org/abs/1802.10419
- github: https://github.com/iboing/CliqueNet
Decoupled Networks
- intro: CVPR 2018 (Spotlight)
- arxiv: https://arxiv.org/abs/1804.08071
Optical Neural Networks
https://arxiv.org/abs/1805.06082
Regularization Learning Networks
- intro: Weizmann Institute of Science
- keywords: Regularization Learning Networks (RLNs), Counterfactual Loss, tabular datasets
- arxiv: https://arxiv.org/abs/1805.06440
Bilinear Attention Networks
https://arxiv.org/abs/1805.07932
Cautious Deep Learning
https://arxiv.org/abs/1805.09460
Perturbative Neural Networks
- intro: CVPR 2018
- intro: We introduce a very simple, yet effective, module called a perturbation layer as an alternative to a convolutional layer
- project page: http://xujuefei.com/pnn.html
- arxiv: https://arxiv.org/abs/1806.01817
Lightweight Probabilistic Deep Networks
- intro: CVPR 2018
- arxiv: https://arxiv.org/abs/1805.11327
Channel Gating Neural Networks
https://arxiv.org/abs/1805.12549
Evenly Cascaded Convolutional Networks
https://arxiv.org/abs/1807.00456
SGAD: Soft-Guided Adaptively-Dropped Neural Network
https://arxiv.org/abs/1807.01430
Explainable Neural Computation via Stack Neural Module Networks
- intro: ECCV 2018
- arxiv: https://arxiv.org/abs/1807.08556
Rank-1 Convolutional Neural Network
https://arxiv.org/abs/1808.04303
Neural Network Encapsulation
- intro: ECCV 2018
- arxiv: https://arxiv.org/abs/1808.03749
Penetrating the Fog: the Path to Efficient CNN Models
https://arxiv.org/abs/1810.04231
A2-Nets: Double Attention Networks
- intro: NIPS 2018
- arxiv: https://arxiv.org/abs/1810.11579
Global Second-order Pooling Neural Networks
https://arxiv.org/abs/1811.12006
ESPNetv2: A Light-weight, Power Efficient, and General Purpose Convolutional Neural Network
- intro: University of Washington & Allen Institute for AI (AI2) & XNOR.AI
- arxiv: https://arxiv.org/abs/1811.11431
- github: https://github.com/sacmehta/ESPNetv2
Kernel Transformer Networks for Compact Spherical Convolution
https://arxiv.org/abs/1812.03115
UAN: Unified Attention Network for Convolutional Neural Networks
https://arxiv.org/abs/1901.05376
One-Class Convolutional Neural Network
- intro: Johns Hopkins University
- arxiv: https://arxiv.org/abs/1901.08688
- github: https://github.com/otkupjnoz/oc-cnn
Selective Kernel Networks
- intro: CVPR 2019
- inrtro: Nanjing University of Science and Technology & Momenta & Nanjing University & Tsinghua University]
- arxiv: https://arxiv.org/abs/1903.06586
- github: https://github.com/implus/SKNet
Universally Slimmable Networks and Improved Training Techniques
- intro: ICLR 2019
- arxiv: https://arxiv.org/abs/1903.05134
Dynamic Slimmable Network
- intro: CVPR 2021 oral
- arxiv: https://arxiv.org/abs/2103.13258
- github: https://github.com/changlin31/DS-Net
Adaptively Connected Neural Networks
- intro: CVPR 2019
- arxiv: https://arxiv.org/abs/1904.03579
- github: https://github.com/wanggrun/Adaptively-Connected-Neural-Networks
Transformable Bottleneck Networks
https://arxiv.org/abs/1904.06458
Pixel-Adaptive Convolutional Neural Networks
- intro: CVPR 2019
- arxiv: https://arxiv.org/abs/1904.05373
Attention Augmented Convolutional Networks
- intro: Google Brain
- arxiv: https://arxiv.org/abs/1904.09925
Spatial Group-wise Enhance: Improving Semantic Feature Learning in Convolutional Networks
EnsembleNet: End-to-End Optimization of Multi-headed Models
- intro: Google AI
- arxiv: https://arxiv.org/abs/1905.09979
MixNet: Mixed Depthwise Convolutional Kernels
- intro: BMVC 2019
- arxiv: https://arxiv.org/abs/1907.09595
- github: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet/mixnet
HarDNet: A Low Memory Traffic Network
- intro: ICCV 2019
- intro: National Tsing Hua University & University of Michigan
- arxiv: https://arxiv.org/abs/1909.00948
Π− nets: Deep Polynomial Neural Networks
- intro: CVPR 2020
- arxiv: https://arxiv.org/abs/2003.03828
Circle Loss: A Unified Perspective of Pair Similarity Optimization
- intro: 1Megvii Inc. & Beihang University & Australian National University & Tsinghua University
- arxiv: https://arxiv.org/abs/2002.10857
Designing Network Design Spaces
- intro: CVPR 2020
- intro: Facebook AI Research (FAIR)
- arxiv: https://arxiv.org/abs/2003.13678
WeightNet: Revisiting the Design Space of Weight Networks
- intro: ECCV 2020
- arxiv: https://arxiv.org/abs/2007.11823
- github: https://github.com/megvii-model/WeightNet
Disentangled Non-Local Neural Networks
https://arxiv.org/abs/2006.06668
Dynamic Neural Networks: A Survey
- intro: Tsinghua University
- arxiv: https://arxiv.org/abs/2102.04906
Convolutions / Filters
Warped Convolutions: Efficient Invariance to Spatial Transformations
Coordinating Filters for Faster Deep Neural Networks
Shift: A Zero FLOP, Zero Parameter Alternative to Spatial Convolutions
- intro: UC Berkeley
- arxiv: https://arxiv.org/abs/1711.08141
Spatially-Adaptive Filter Units for Deep Neural Networks
- intro: University of Ljubljana & University of Birmingham
- arxiv: https://arxiv.org/abs/1711.11473
clcNet: Improving the Efficiency of Convolutional Neural Network using Channel Local Convolutions
https://arxiv.org/abs/1712.06145
DCFNet: Deep Neural Network with Decomposed Convolutional Filters
https://arxiv.org/abs/1802.04145
Fast End-to-End Trainable Guided Filter
- intro: CVPR 2018
- project page: http://wuhuikai.me/DeepGuidedFilterProject/
- gtihub(official, PyTorch): https://github.com/wuhuikai/DeepGuidedFilter
Diagonalwise Refactorization: An Efficient Training Method for Depthwise Convolutions
- arxiv: https://arxiv.org/abs/1803.09926
- github: https://github.com/clavichord93/diagonalwise-refactorization-tensorflow
Use of symmetric kernels for convolutional neural networks
- intro: ICDSIAI 2018
- arxiv: https://arxiv.org/abs/1805.09421
EasyConvPooling: Random Pooling with Easy Convolution for Accelerating Training and Testing
https://arxiv.org/abs/1806.01729
Targeted Kernel Networks: Faster Convolutions with Attentive Regularization
https://arxiv.org/abs/1806.00523
An Intriguing Failing of Convolutional Neural Networks and the CoordConv Solution
- intro: NeurIPS 2018
- intro: Uber AI Labs & Uber Technologies
- arxiv: https://arxiv.org/abs/1807.03247
- github: https://github.com/uber-research/CoordConv
- youtube: https://www.youtube.com/watch?v=8yFQc6elePA
Network Decoupling: From Regular to Depthwise Separable Convolutions
https://arxiv.org/abs/1808.05517
Partial Convolution based Padding
- intro: NVIDIA Corporation
- arxiv; https://arxiv.org/abs/1811.11718
- github: https://github.com/NVIDIA/partialconv
DSConv: Efficient Convolution Operator
https://arxiv.org/abs/1901.01928
CircConv: A Structured Convolution with Low Complexity
- intro: AAAI 2019
- arxiv: https://arxiv.org/abs/1902.11268
Accelerating Large-Kernel Convolution Using Summed-Area Tables
- intro: Princeton University
- arxiv: https://arxiv.org/abs/1906.11367
Mapped Convolutions
- intro: University of North Carolina at Chapel Hill
- arxiv: https://arxiv.org/abs/1906.11096
Universal Pooling – A New Pooling Method for Convolutional Neural Networks
https://arxiv.org/abs/1907.11440
Dilated Point Convolutions: On the Receptive Field of Point Convolutions
https://arxiv.org/abs/1907.12046
LIP: Local Importance-based Pooling
- intro: ICCV 2019
- arxiv: https://arxiv.org/abs/1908.04156
Deep Generalized Max Pooling
- intro: ICDAR
- arxiv: https://arxiv.org/abs/1908.05040
MixConv: Mixed Depthwise Convolutional Kernels
- intro:BMVC 2019
- arxiv: https://arxiv.org/abs/1907.09595
- github: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet/mixnet
Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation
- intro: UC Berkeley & USTC & MSRA
- arxiv: https://arxiv.org/abs/1910.02940
Dynamic Convolution: Attention over Convolution Kernels
- intro: CVPR 2020 oral
- intro: Microsoft
- arxiv: https://arxiv.org/abs/1912.03458
Pyramidal Convolution: Rethinking Convolutional Neural Networks for Visual Recognition
- arxiv: https://arxiv.org/abs/2006.11538
- github: https://github.com/iduta/pyconv
- gihtub: https://github.com/iduta/pyconvsegnet
Highway Networks
Highway Networks
- intro: ICML 2015 Deep Learning workshop
- intro: shortcut connections with gating functions. These gates are data-dependent and have parameters
- arxiv: http://arxiv.org/abs/1505.00387
- github(PyTorch): https://github.com/analvikingur/pytorch_Highway
Highway Networks with TensorFlow
Very Deep Learning with Highway Networks
- homepage(papers+code+FAQ): http://people.idsia.ch/~rupesh/very_deep_learning/
Training Very Deep Networks
- intro: Extends Highway Networks
- project page: http://people.idsia.ch/~rupesh/very_deep_learning/
- arxiv: http://arxiv.org/abs/1507.06228
Spatial Transformer Networks
Spatial Transformer Networks
- intro: NIPS 2015
- arxiv: http://arxiv.org/abs/1506.02025
- gitxiv: http://gitxiv.com/posts/5WTXTLuEA4Hd8W84G/spatial-transformer-networks
- github: https://github.com/daerduoCarey/SpatialTransformerLayer
- github: https://github.com/qassemoquab/stnbhwd
- github: https://github.com/skaae/transformer_network
- github(Caffe): https://github.com/happynear/SpatialTransformerLayer
- github: https://github.com/daviddao/spatial-transformer-tensorflow
- caffe-issue: https://github.com/BVLC/caffe/issues/3114
- code: https://lasagne.readthedocs.org/en/latest/modules/layers/special.html#lasagne.layers.TransformerLayer
- ipn(Lasagne): http://nbviewer.jupyter.org/github/Lasagne/Recipes/blob/master/examples/spatial_transformer_network.ipynb
- notes: https://www.evernote.com/shard/s189/sh/ad8a38de-9e98-4e06-b09e-574bd62893ff/32f72798c095dd7672f4cb017a32d9b4
- youtube: https://www.youtube.com/watch?v=6NOQC_fl1hQ
The power of Spatial Transformer Networks
- blog: http://torch.ch/blog/2015/09/07/spatial_transformers.html
- github: https://github.com/moodstocks/gtsrb.torch
Recurrent Spatial Transformer Networks
Deep Learning Paper Implementations: Spatial Transformer Networks - Part I
- blog: https://kevinzakka.github.io/2017/01/10/stn-part1/
- github: https://github.com/kevinzakka/blog-code/tree/master/spatial_transformer
Top-down Flow Transformer Networks
https://arxiv.org/abs/1712.02400
Non-Parametric Transformation Networks
- intro: CMU
- arxiv: https://arxiv.org/abs/1801.04520
Hierarchical Spatial Transformer Network
https://arxiv.org/abs/1801.09467
Spatial Transformer Introspective Neural Network
- intro: Johns Hopkins University & Shanghai University
- arxiv: https://arxiv.org/abs/1805.06447
DeSTNet: Densely Fused Spatial Transformer Networks
- intro: BMVC 2018
- arxiv: https://arxiv.org/abs/1807.04050
MIST: Multiple Instance Spatial Transformer Network
https://arxiv.org/abs/1811.10725
FractalNet
FractalNet: Ultra-Deep Neural Networks without Residuals
- project: http://people.cs.uchicago.edu/~larsson/fractalnet/
- arxiv: http://arxiv.org/abs/1605.07648
- github: https://github.com/gustavla/fractalnet
- github: https://github.com/edgelord/FractalNet
- github(Keras): https://github.com/snf/keras-fractalnet
Generative Models
Max-margin Deep Generative Models
- intro: NIPS 2015
- arxiv: http://arxiv.org/abs/1504.06787
- github: https://github.com/zhenxuan00/mmdgm
Discriminative Regularization for Generative Models
Auxiliary Deep Generative Models
- arxiv: http://arxiv.org/abs/1602.05473
- github: https://github.com/larsmaaloee/auxiliary-deep-generative-models
Sampling Generative Networks: Notes on a Few Effective Techniques
Conditional Image Synthesis With Auxiliary Classifier GANs
- arxiv: https://arxiv.org/abs/1610.09585
- github: https://github.com/buriburisuri/ac-gan
- github(Keras): https://github.com/lukedeo/keras-acgan
On the Quantitative Analysis of Decoder-Based Generative Models
- intro: University of Toronto & OpenAI & CMU
- arxiv: https://arxiv.org/abs/1611.04273
- github: https://github.com/tonywu95/eval_gen
Boosted Generative Models
An Architecture for Deep, Hierarchical Generative Models
- intro: NIPS 2016
- arxiv: https://arxiv.org/abs/1612.04739
- github: https://github.com/Philip-Bachman/MatNets-NIPS
Deep Learning and Hierarchal Generative Models
- intro: NIPS 2016. MIT
- arxiv: https://arxiv.org/abs/1612.09057
Probabilistic Torch
- intro: Probabilistic Torch is library for deep generative models that extends PyTorch
- github: https://github.com/probtorch/probtorch
Tutorial on Deep Generative Models
- intro: UAI 2017 Tutorial: Shakir Mohamed & Danilo Rezende (DeepMind)
- youtube: https://www.youtube.com/watch?v=JrO5fSskISY
- mirror: https://www.bilibili.com/video/av16428277/
- slides: http://www.shakirm.com/slides/DeepGenModelsTutorial.pdf
A Note on the Inception Score
- intro: Stanford University
- arxiv: https://arxiv.org/abs/1801.01973
Gradient Layer: Enhancing the Convergence of Adversarial Training for Generative Models
- intro: AISTATS 2018. The University of Tokyo
- arxiv: https://arxiv.org/abs/1801.02227
Batch Normalization in the final layer of generative networks
https://arxiv.org/abs/1805.07389
Deep Structured Generative Models
- intro: Tsinghua University
- arxiv: https://arxiv.org/abs/1807.03877
VFunc: a Deep Generative Model for Functions
- intro: ICML 2018 workshop on Prediction and Generative Modeling in Reinforcement Learning. Microsoft Research & McGill University
- arxiv: https://arxiv.org/abs/1807.04106
Deep Learning and Robots
Robot Learning Manipulation Action Plans by “Watching” Unconstrained Videos from the World Wide Web
- intro: AAAI 2015
- paper: http://www.umiacs.umd.edu/~yzyang/paper/YouCookMani_CameraReady.pdf
- author page: http://www.umiacs.umd.edu/~yzyang/
End-to-End Training of Deep Visuomotor Policies
Comment on Open AI’s Efforts to Robot Learning
The Curious Robot: Learning Visual Representations via Physical Interactions
How to build a robot that “sees” with $100 and TensorFlow
Deep Visual Foresight for Planning Robot Motion
- project page: https://sites.google.com/site/brainrobotdata/
- arxiv: https://arxiv.org/abs/1610.00696
- video: https://sites.google.com/site/robotforesight/
Sim-to-Real Robot Learning from Pixels with Progressive Nets
- intro: Google DeepMind
- arxiv: https://arxiv.org/abs/1610.04286
Towards Lifelong Self-Supervision: A Deep Learning Direction for Robotics
A Differentiable Physics Engine for Deep Learning in Robotics
Deep-learning in Mobile Robotics - from Perception to Control Systems: A Survey on Why and Why not
- intro: City University of Hong Kong & Hong Kong University of Science and Technology
- arxiv: https://arxiv.org/abs/1612.07139
Deep Robotic Learning
- intro: https://simons.berkeley.edu/talks/sergey-levine-01-24-2017-1
- youtube: https://www.youtube.com/watch?v=jtjW5Pye_44
Deep Learning in Robotics: A Review of Recent Research
https://arxiv.org/abs/1707.07217
Deep Learning for Robotics
- intro: by Pieter Abbeel
- video: https://www.facebook.com/nipsfoundation/videos/1554594181298482/
- mirror: https://www.bilibili.com/video/av17078186/
- slides: https://www.dropbox.com/s/4fhczb9cxkuqalf/2017_11_xx_BARS-Abbeel.pdf?dl=0
DroNet: Learning to Fly by Driving
- project page: http://rpg.ifi.uzh.ch/dronet.html
- paper: http://rpg.ifi.uzh.ch/docs/RAL18_Loquercio.pdf
- github: https://github.com/uzh-rpg/rpg_public_dronet
A Survey on Deep Learning Methods for Robot Vision
https://arxiv.org/abs/1803.10862
Deep Learning on Mobile / Embedded Devices
Convolutional neural networks on the iPhone with VGGNet
- blog: http://matthijshollemans.com/2016/08/30/vggnet-convolutional-neural-network-iphone/
- github: https://github.com/hollance/VGGNet-Metal
TensorFlow for Mobile Poets
The Convolutional Neural Network(CNN) for Android
- intro: CnnForAndroid:A Classification Project using Convolutional Neural Network(CNN) in Android platform。It also support Caffe Model
- github: https://github.com/zhangqianhui/CnnForAndroid
TensorFlow on Android
Experimenting with TensorFlow on Android
- part 1: https://medium.com/@mgazar/experimenting-with-tensorflow-on-android-pt-1-362683b31838#.5gbp2d4st
- part 2: https://medium.com/@mgazar/experimenting-with-tensorflow-on-android-part-2-12f3dc294eaf#.2gx3o65f5
- github: https://github.com/MostafaGazar/tensorflow
XNOR.ai frees AI from the prison of the supercomputer
Embedded and mobile deep learning research resources
https://github.com/csarron/emdl
Modeling the Resource Requirements of Convolutional Neural Networks on Mobile Devices
https://arxiv.org/abs/1709.09118
Benchmarks
Deep Learning’s Accuracy
Benchmarks for popular CNN models
- intro: Benchmarks for popular convolutional neural network models on CPU and different GPUs, with and without cuDNN.
- github: https://github.com/jcjohnson/cnn-benchmarks
Deep Learning Benchmarks
http://add-for.com/deep-learning-benchmarks/
cudnn-rnn-benchmarks
Papers
Reweighted Wake-Sleep
Probabilistic Backpropagation for Scalable Learning of Bayesian Neural Networks
- paper: http://arxiv.org/abs/1502.05336
- github: https://github.com/HIPS/Probabilistic-Backpropagation
Deeply-Supervised Nets
Deep learning
- intro: Nature 2015
- author: Yann LeCun, Yoshua Bengio & Geoffrey Hinton
- paper: http://www.cs.toronto.edu/~hinton/absps/NatureDeepReview.pdf
On the Expressive Power of Deep Learning: A Tensor Analysis
Understanding and Predicting Image Memorability at a Large Scale
- intro: MIT. ICCV 2015
- homepage: http://memorability.csail.mit.edu/
- paper: https://people.csail.mit.edu/khosla/papers/iccv2015_khosla.pdf
- code: http://memorability.csail.mit.edu/download.html
- reviews: http://petapixel.com/2015/12/18/how-memorable-are-times-top-10-photos-of-2015-to-a-computer/
Towards Open Set Deep Networks
Structured Prediction Energy Networks
- intro: ICML 2016. SPEN
- arxiv: http://arxiv.org/abs/1511.06350
- github: https://github.com/davidBelanger/SPEN
Deep Neural Networks predict Hierarchical Spatio-temporal Cortical Dynamics of Human Visual Object Recognition
Recent Advances in Convolutional Neural Networks
Understanding Deep Convolutional Networks
DeepCare: A Deep Dynamic Memory Model for Predictive Medicine
Exploiting Cyclic Symmetry in Convolutional Neural Networks
- intro: ICML 2016
- arxiv: http://arxiv.org/abs/1602.02660
- github(Winning solution for the National Data Science Bowl competition on Kaggle (plankton classification)): https://github.com/benanne/kaggle-ndsb
- ref(use Cyclic pooling): http://benanne.github.io/2015/03/17/plankton.html
Cross-dimensional Weighting for Aggregated Deep Convolutional Features
- arxiv: http://arxiv.org/abs/1512.04065
- github: https://github.com/yahoo/crow
Understanding Visual Concepts with Continuation Learning
- project page: http://willwhitney.github.io/understanding-visual-concepts/
- arxiv: http://arxiv.org/abs/1602.06822
- github: https://github.com/willwhitney/understanding-visual-concepts
Learning Efficient Algorithms with Hierarchical Attentive Memory
- arxiv: http://arxiv.org/abs/1602.03218
- github: https://github.com/Smerity/tf-ham
DrMAD: Distilling Reverse-Mode Automatic Differentiation for Optimizing Hyperparameters of Deep Neural Networks
Do Deep Convolutional Nets Really Need to be Deep (Or Even Convolutional)?
- arxiv: http://arxiv.org/abs/1603.05691
- review: http://www.erogol.com/paper-review-deep-convolutional-nets-really-need-deep-even-convolutional/
Harnessing Deep Neural Networks with Logic Rules
Degrees of Freedom in Deep Neural Networks
Deep Networks with Stochastic Depth
- arxiv: http://arxiv.org/abs/1603.09382
- github: https://github.com/yueatsprograms/Stochastic_Depth
- notes(“Stochastic Depth Networks will Become the New Normal”): http://deliprao.com/archives/134
- github: https://github.com/dblN/stochastic_depth_keras
- github: https://github.com/yasunorikudo/chainer-ResDrop
- review: https://medium.com/@tim_nth/review-deep-networks-with-stochastic-depth-51bd53acfe72
LIFT: Learned Invariant Feature Transform
- intro: ECCV 2016
- arxiv: http://arxiv.org/abs/1603.09114
- github(official): https://github.com/cvlab-epfl/LIFT
Bridging the Gaps Between Residual Learning, Recurrent Neural Networks and Visual Cortex
- arxiv: https://arxiv.org/abs/1604.03640
- slides: http://prlab.tudelft.nl/sites/default/files/rnnResnetCortex.pdf
Understanding How Image Quality Affects Deep Neural Networks
- arxiv: http://arxiv.org/abs/1604.04004
- reddit: https://www.reddit.com/r/MachineLearning/comments/4exk3u/dcnns_are_more_sensitive_to_blur_and_noise_than/
Deep Embedding for Spatial Role Labeling
- arxiv: http://arxiv.org/abs/1603.08474
- github: https://github.com/oswaldoludwig/visually-informed-embedding-of-word-VIEW-
Unreasonable Effectiveness of Learning Neural Nets: Accessible States and Robust Ensembles
Learning Deep Representation for Imbalanced Classification
- intro: CVPR 2016
- keywords: Deep Learning Large Margin Local Embedding (LMLE)
- project page: http://mmlab.ie.cuhk.edu.hk/projects/LMLE.html
- paper: http://personal.ie.cuhk.edu.hk/~ccloy/files/cvpr_2016_imbalanced.pdf
- code: http://mmlab.ie.cuhk.edu.hk/projects/LMLE/lmle_code.zip
Newtonian Image Understanding: Unfolding the Dynamics of Objects in Static Images
- homepage: http://allenai.org/plato/newtonian-understanding/
- arxiv: http://arxiv.org/abs/1511.04048
- github: https://github.com/roozbehm/newtonian
DeepMath - Deep Sequence Models for Premise Selection
Convolutional Neural Networks Analyzed via Convolutional Sparse Coding
Systematic evaluation of CNN advances on the ImageNet
Why does deep and cheap learning work so well?
- intro: Harvard and MIT
- arxiv: http://arxiv.org/abs/1608.08225
- review: https://www.technologyreview.com/s/602344/the-extraordinary-link-between-deep-neural-networks-and-the-nature-of-the-universe/
A scalable convolutional neural network for task-specified scenarios via knowledge distillation
Alternating Back-Propagation for Generator Network
- project page(code+data): http://www.stat.ucla.edu/~ywu/ABP/main.html
- paper: http://www.stat.ucla.edu/~ywu/ABP/doc/arXivABP.pdf
A Novel Representation of Neural Networks
Optimization of Convolutional Neural Network using Microcanonical Annealing Algorithm
- intro: IEEE ICACSIS 2016
- arxiv: https://arxiv.org/abs/1610.02306
Uncertainty in Deep Learning
- intro: PhD Thesis. Cambridge Machine Learning Group
- blog: http://mlg.eng.cam.ac.uk/yarin/blog_2248.html
- thesis: http://mlg.eng.cam.ac.uk/yarin/thesis/thesis.pdf
Deep Convolutional Neural Network Design Patterns
Extensions and Limitations of the Neural GPU
Neural Functional Programming
Deep Information Propagation
Compressed Learning: A Deep Neural Network Approach
A backward pass through a CNN using a generative model of its activations
Understanding deep learning requires rethinking generalization
- intro: ICLR 2017 best paper. MIT & Google Brain & UC Berkeley & Google DeepMind
- arxiv: https://arxiv.org/abs/1611.03530
- example code: https://github.com/pluskid/fitting-random-labels
- notes: https://theneuralperspective.com/2017/01/24/understanding-deep-learning-requires-rethinking-generalization/
Learning the Number of Neurons in Deep Networks
- intro: NIPS 2016
- arxiv: https://arxiv.org/abs/1611.06321
Survey of Expressivity in Deep Neural Networks
- intro: Presented at NIPS 2016 Workshop on Interpretable Machine Learning in Complex Systems
- intro: Google Brain & Cornell University & Stanford University
- arxiv: https://arxiv.org/abs/1611.08083
Designing Neural Network Architectures using Reinforcement Learning
- intro: MIT
- project page: https://bowenbaker.github.io/metaqnn/
- arxiv: https://arxiv.org/abs/1611.02167
Towards Robust Deep Neural Networks with BANG
- intro: University of Colorado
- arxiv: https://arxiv.org/abs/1612.00138
Deep Quantization: Encoding Convolutional Activations with Deep Generative Model
- intro: University of Science and Technology of China & MSR
- arxiv: https://arxiv.org/abs/1611.09502
A Probabilistic Theory of Deep Learning
A Probabilistic Framework for Deep Learning
- intro: Rice University
- arxiv: https://arxiv.org/abs/1612.01936
Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer
- arxiv: https://arxiv.org/abs/1612.03928
- github(PyTorch): https://github.com/szagoruyko/attention-transfer
Risk versus Uncertainty in Deep Learning: Bayes, Bootstrap and the Dangers of Dropout
- intro: Google Deepmind
- paper: http://bayesiandeeplearning.org/papers/BDL_4.pdf
Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
- intro: Google Brain & Jagiellonian University
- keywords: Sparsely-Gated Mixture-of-Experts layer (MoE), language modeling and machine translation
- arxiv: https://arxiv.org/abs/1701.06538
- reddit: https://www.reddit.com/r/MachineLearning/comments/5pud72/research_outrageously_large_neural_networks_the/
Deep Network Guided Proof Search
- intro: Google Research & University of Innsbruck
- arxiv: https://arxiv.org/abs/1701.06972
PathNet: Evolution Channels Gradient Descent in Super Neural Networks
- intro: Google DeepMind & Google Brain
- arxiv: https://arxiv.org/abs/1701.08734
- notes: https://medium.com/intuitionmachine/pathnet-a-modular-deep-learning-architecture-for-agi-5302fcf53273#.8f0o6w3en
Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networks
The Power of Sparsity in Convolutional Neural Networks
Learning across scales - A multiscale method for Convolution Neural Networks
Stacking-based Deep Neural Network: Deep Analytic Network on Convolutional Spectral Histogram Features
A Compositional Object-Based Approach to Learning Physical Dynamics
- intro: ICLR 2017. Neural Physics Engine
- paper: https://openreview.net/pdf?id=Bkab5dqxe
- github: https://github.com/mbchang/dynamics
Genetic CNN
- arxiv: https://arxiv.org/abs/1703.01513
- github(Tensorflow): https://github.com/aqibsaeed/Genetic-CNN
Deep Sets
- intro: Amazon Web Services & CMU
- keywords: statistic estimation, point cloud classification, set expansion, and image tagging
- arxiv: https://arxiv.org/abs/1703.06114
Multiscale Hierarchical Convolutional Networks
Deep Neural Networks Do Not Recognize Negative Images
https://arxiv.org/abs/1703.06857
Failures of Deep Learning
Multi-Scale Dense Convolutional Networks for Efficient Prediction
- intro: Cornell University & Tsinghua University & Fudan University & Facebook AI Research
- arxiv: https://arxiv.org/abs/1703.09844
- github: https://github.com/gaohuang/MSDNet
Scaling the Scattering Transform: Deep Hybrid Networks
- arxiv: https://arxiv.org/abs/1703.08961
- github: https://github.com/edouardoyallon/scalingscattering
- github(CuPy/PyTorch): https://github.com/edouardoyallon/pyscatwave
Deep Learning is Robust to Massive Label Noise
https://arxiv.org/abs/1705.10694
Input Fast-Forwarding for Better Deep Learning
- intro: ICIAR 2017
- keywords: Fast-Forward Network (FFNet)
- arxiv: https://arxiv.org/abs/1705.08479
Deep Mutual Learning
- intro: CVPR 2018
- keywords: deep mutual learning (DML)
- arxiv: https://arxiv.org/abs/1706.00384
- github(official, TensorFlow): https://github.com/YingZhangDUT/Deep-Mutual-Learning
Automated Problem Identification: Regression vs Classification via Evolutionary Deep Networks
- intro: University of Cape Town
- arxiv: https://arxiv.org/abs/1707.00703
Revisiting Unreasonable Effectiveness of Data in Deep Learning Era
- intro: Google Research & CMU
- arxiv: https://arxiv.org/abs/1707.02968
- blog: https://research.googleblog.com/2017/07/revisiting-unreasonable-effectiveness.html
Deep Layer Aggregation
- intro: UC Berkeley
- arxiv: https://arxiv.org/abs/1707.06484
Improving Robustness of Feature Representations to Image Deformations using Powered Convolution in CNNs
https://arxiv.org/abs/1707.07830
Learning uncertainty in regression tasks by deep neural networks
- intro: Free University of Berlin
- arxiv: https://arxiv.org/abs/1707.07287
Generalizing the Convolution Operator in Convolutional Neural Networks
https://arxiv.org/abs/1707.09864
Convolution with Logarithmic Filter Groups for Efficient Shallow CNN
https://arxiv.org/abs/1707.09855
Deep Multi-View Learning with Stochastic Decorrelation Loss
https://arxiv.org/abs/1707.09669
Take it in your stride: Do we need striding in CNNs?
https://arxiv.org/abs/1712.02502
Security Risks in Deep Learning Implementation
- intro: Qihoo 360 Security Research Lab & University of Georgia & University of Virginia
- arxiv: https://arxiv.org/abs/1711.11008
Online Learning with Gated Linear Networks
- intro: DeepMind
- arxiv: https://arxiv.org/abs/1712.01897
On the Information Bottleneck Theory of Deep Learning
https://openreview.net/forum?id=ry_WPG-A-¬eId=ry_WPG-A
The Unreasonable Effectiveness of Deep Features as a Perceptual Metric
- project page: https://richzhang.github.io/PerceptualSimilarity/
- arxiv: https://arxiv.org/abs/1801.03924
- github: https://github.com//richzhang/PerceptualSimilarity
Less is More: Culling the Training Set to Improve Robustness of Deep Neural Networks
- intro: University of California, Davis
- arxiv: https://arxiv.org/abs/1801.02850
Towards an Understanding of Neural Networks in Natural-Image Spaces
https://arxiv.org/abs/1801.09097
Deep Private-Feature Extraction
https://arxiv.org/abs/1802.03151
Not All Samples Are Created Equal: Deep Learning with Importance Sampling
- intro: Idiap Research Institute
- arxiv: https://arxiv.org/abs/1803.00942
Label Refinery: Improving ImageNet Classification through Label Progression
- intro: Using a Label Refinery improves the state-of-the-art top-1 accuracy of (1) AlexNet from 59.3 to 67.2, (2) MobileNet from 70.6 to 73.39, (3) MobileNet-0.25 from 50.6 to 55.59, (4) VGG19 from 72.7 to 75.46, and (5) Darknet19 from 72.9 to 74.47.
- intro: XNOR AI, University of Washington, Allen AI
- arxiv: https://arxiv.org/abs/1805.02641
- github: https://github.com/hessamb/label-refinery
How Many Samples are Needed to Learn a Convolutional Neural Network?
https://arxiv.org/abs/1805.07883
VisualBackProp for learning using privileged information with CNNs
https://arxiv.org/abs/1805.09474
BAM: Bottleneck Attention Module
- intro: BMVC 2018 (oral). Lunit Inc. & Adobe Research
- arxiv: https://arxiv.org/abs/1807.06514
CBAM: Convolutional Block Attention Module
- intro: ECCV 2018. Lunit Inc. & Adobe Research
- arxiv: https://arxiv.org/abs/1807.06521
Scale equivariance in CNNs with vector fields
- intro: ICML/FAIM 2018 workshop on Towards learning with limited labels: Equivariance, Invariance, and Beyond (oral presentation)
- arxiv: https://arxiv.org/abs/1807.11783
Downsampling leads to Image Memorization in Convolutional Autoencoders
https://arxiv.org/abs/1810.10333
Do Normalization Layers in a Deep ConvNet Really Need to Be Distinct?
https://arxiv.org/abs/1811.07727
Are All Training Examples Created Equal? An Empirical Study
https://arxiv.org/abs/1811.12569
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness
https://arxiv.org/abs/1811.12231
DeepFashion2: A Versatile Benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images
- intro: The Chinese University of Hong Kong & SenseTime Research
- keywords: Match R-CNN
- arxiv: https://arxiv.org/abs/1901.07973
A Comprehensive Overhaul of Feature Distillation
https://arxiv.org/abs/1904.01866
Mesh R-CNN
- intro: Facebook AI Research (FAIR)
- arxiv: https://arxiv.org/abs/1906.02739
ViP: Virtual Pooling for Accelerating CNN-based Image Classification and Object Detection
https://arxiv.org/abs/1906.07912
VarGNet: Variable Group Convolutional Neural Network for Efficient Embedded Computing
- intro: Horizon Robotics
- arxiv: https://arxiv.org/abs/1907.05653
Anchor Loss: Modulating Loss Scale based on Prediction Difficulty
- intro: ICCV 2019 oral
- arxiv: https://arxiv.org/abs/1909.11155
Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem
- intro: CVPR 2019 oral
- arxiv: https://arxiv.org/abs/1812.05720
- github: https://github.com/max-andr/relu_networks_overconfident
Feature Space Augmentation for Long-Tailed Data
- intro: ECCV 2020
- arxiv: https://arxiv.org/abs/2008.03673
Tutorials and Surveys
A Survey: Time Travel in Deep Learning Space: An Introduction to Deep Learning Models and How Deep Learning Models Evolved from the Initial Ideas
On the Origin of Deep Learning
- intro: CMU. 70 pages, 200 references
- arxiv: https://arxiv.org/abs/1702.07800
Efficient Processing of Deep Neural Networks: A Tutorial and Survey
- intro: MIT
- arxiv: https://arxiv.org/abs/1703.09039
The History Began from AlexNet: A Comprehensive Survey on Deep Learning Approaches
{https://arxiv.org/abs/1803.01164}(https://arxiv.org/abs/1803.01164)
Mathematics of Deep Learning
A Mathematical Theory of Deep Convolutional Neural Networks for Feature Extraction
Mathematics of Deep Learning
- intro: Johns Hopkins University & New York University & Tel-Aviv University & University of California, Los Angeles
- arxiv: https://arxiv.org/abs/1712.04741
Local Minima
Local minima in training of deep networks
- intro: DeepMind
- arxiv: https://arxiv.org/abs/1611.06310
Deep linear neural networks with arbitrary loss: All local minima are global
- intro: CMU & University of Southern California & Facebook Artificial Intelligence Research
- arxiv: https://arxiv.org/abs/1712.00779
Gradient Descent Learns One-hidden-layer CNN: Don’t be Afraid of Spurious Local Minima
- intro: Loyola Marymount University & California State University
- arxiv: https://arxiv.org/abs/1712.01473
CNNs are Globally Optimal Given Multi-Layer Support
- intro: CMU
- arxiv: https://arxiv.org/abs/1712.02501
Spurious Local Minima are Common in Two-Layer ReLU Neural Networks
https://arxiv.org/abs/1712.08968
Dive Into CNN
Structured Receptive Fields in CNNs
How ConvNets model Non-linear Transformations
Separable Convolutions / Grouped Convolutions
Factorized Convolutional Neural Networks
Design of Efficient Convolutional Layers using Single Intra-channel Convolution, Topological Subdivisioning and Spatial “Bottleneck” Structure
XSepConv: Extremely Separated Convolution
- intro: Tsinghua University & University College London
- arxiv: https://arxiv.org/abs/2002.12046
STDP
A biological gradient descent for prediction through a combination of STDP and homeostatic plasticity
An objective function for STDP
Towards a Biologically Plausible Backprop
Target Propagation
How Auto-Encoders Could Provide Credit Assignment in Deep Networks via Target Propagation
Difference Target Propagation
- arxiv: http://arxiv.org/abs/1412.7525
- github: https://github.com/donghyunlee/dtp
Zero Shot Learning
Learning a Deep Embedding Model for Zero-Shot Learning
Zero-Shot (Deep) Learning
https://amundtveit.com/2016/11/18/zero-shot-deep-learning/
Zero-shot learning experiments by deep learning.
https://github.com/Elyorcv/zsl-deep-learning
Zero-Shot Learning - The Good, the Bad and the Ugly
- intro: CVPR 2017
- arxiv: https://arxiv.org/abs/1703.04394
Semantic Autoencoder for Zero-Shot Learning
- intro: CVPR 2017
- project page: https://elyorcv.github.io/projects/sae
- arxiv: https://arxiv.org/abs/1704.08345
- github: https://github.com/Elyorcv/SAE
Zero-Shot Learning via Category-Specific Visual-Semantic Mapping
https://arxiv.org/abs/1711.06167
Zero-Shot Learning via Class-Conditioned Deep Generative Models
- intro: AAAI 2018
- arxiv: https://arxiv.org/abs/1711.05820
Feature Generating Networks for Zero-Shot Learning
https://arxiv.org/abs/1712.00981
Zero-Shot Visual Recognition using Semantics-Preserving Adversarial Embedding Network
https://arxiv.org/abs/1712.01928
Combining Deep Universal Features, Semantic Attributes, and Hierarchical Classification for Zero-Shot Learning
- intro: extension to work published in conference proceedings of 2017 IAPR MVA Conference
- arxiv: https://arxiv.org/abs/1712.03151
Multi-Context Label Embedding
- keywords: Multi-Context Label Embedding (MCLE)
- arxiv: https://arxiv.org/abs/1805.01199
Incremental Learning
iCaRL: Incremental Classifier and Representation Learning
FearNet: Brain-Inspired Model for Incremental Learning
https://arxiv.org/abs/1711.10563
Incremental Learning in Deep Convolutional Neural Networks Using Partial Network Sharing
- intro: Purdue University
- arxiv: https://arxiv.org/abs/1712.02719
Incremental Classifier Learning with Generative Adversarial Networks
https://arxiv.org/abs/1802.00853
Learn the new, keep the old: Extending pretrained models with new anatomy and images
- intro: MICCAI 2018
- arxiv: https://arxiv.org/abs/1806.00265
Ensemble Deep Learning
Convolutional Neural Fabrics
- intro: NIPS 2016
- arxiv: http://arxiv.org/abs/1606.02492
- github: https://github.com/shreyassaxena/convolutional-neural-fabrics
Stochastic Multiple Choice Learning for Training Diverse Deep Ensembles
- arxiv: https://arxiv.org/abs/1606.07839
- youtube: https://www.youtube.com/watch?v=KjUfMtZjyfg&feature=youtu.be
Snapshot Ensembles: Train 1, Get M for Free
- paper: http://openreview.net/pdf?id=BJYwwY9ll
- github(Torch): https://github.com/gaohuang/SnapshotEnsemble
- github: https://github.com/titu1994/Snapshot-Ensembles
Ensemble Deep Learning
Domain Adaptation
Adversarial Discriminative Domain Adaptation
- intro: UC Berkeley & Stanford University & Boston University
- arxiv: https://arxiv.org/abs/1702.05464
- github: https://github.com//corenel/pytorch-adda
Parameter Reference Loss for Unsupervised Domain Adaptation
https://arxiv.org/abs/1711.07170
Residual Parameter Transfer for Deep Domain Adaptation
https://arxiv.org/abs/1711.07714
Adversarial Feature Augmentation for Unsupervised Domain Adaptation
https://arxiv.org/abs/1711.08561
Image to Image Translation for Domain Adaptation
https://arxiv.org/abs/1712.00479
Incremental Adversarial Domain Adaptation
https://arxiv.org/abs/1712.07436
Deep Visual Domain Adaptation: A Survey
https://arxiv.org/abs/1802.03601
Unsupervised Domain Adaptation: A Multi-task Learning-based Method
https://arxiv.org/abs/1803.09208
Importance Weighted Adversarial Nets for Partial Domain Adaptation
https://arxiv.org/abs/1803.09210
Open Set Domain Adaptation by Backpropagation
https://arxiv.org/abs/1804.10427
Learning Sampling Policies for Domain Adaptation
- intro: CMU
- arxiv: https://arxiv.org/abs/1805.07641
Multi-Adversarial Domain Adaptation
- intro: AAAI 2018 Oral.
- arxiv: https://arxiv.org/abs/1809.02176
Unsupervised Domain Adaptation: An Adaptive Feature Norm Approach
- intro: Sun Yat-sen University
- arxiv: https://arxiv.org/abs/1811.07456
- github: https://github.com/jihanyang/AFN/
Multi-source Distilling Domain Adaptation
- intro: AAAI 2020
- arxiv: https://arxiv.org/abs/1911.11554
- github: https://github.com/daoyuan98/MDDA
awsome-domain-adaptation
https://github.com/zhaoxin94/awsome-domain-adaptation
Embedding
Learning Deep Embeddings with Histogram Loss
- intro: NIPS 2016
- arxiv: https://arxiv.org/abs/1611.00822
Full-Network Embedding in a Multimodal Embedding Pipeline
https://arxiv.org/abs/1707.09872
Clustering-driven Deep Embedding with Pairwise Constraints
https://arxiv.org/abs/1803.08457
Deep Mixture of Experts via Shallow Embedding
https://arxiv.org/abs/1806.01531
Learning to Learn from Web Data through Deep Semantic Embeddings
- intro: ECCV MULA Workshop 2018
- arxiv: https://arxiv.org/abs/1808.06368
Heated-Up Softmax Embedding
https://arxiv.org/abs/1809.04157
Virtual Class Enhanced Discriminative Embedding Learning
- intro: NeurIPS 2018
- arxiv: https://arxiv.org/abs/1811.12611
Regression
A Comprehensive Analysis of Deep Regression
https://arxiv.org/abs/1803.08450
Neural Motifs: Scene Graph Parsing with Global Context
- intro: CVPR 2018. University of Washington
- project page: http://rowanzellers.com/neuralmotifs/
- arxiv: https://arxiv.org/abs/1711.06640
- github: https://github.com/rowanz/neural-motifs
- demo: https://rowanzellers.com/scenegraph2/
CapsNets
Dynamic Routing Between Capsules
- intro: Sara Sabour, Nicholas Frosst, Geoffrey E Hinton
- intro: Google Brain, Toronto
- arxiv: https://arxiv.org/abs/1710.09829
- github(official, Tensorflow): https://github.com/Sarasra/models/tree/master/research/capsules
Capsule Networks (CapsNets) – Tutorial
- youtube: https://www.youtube.com/watch?v=pPN8d0E3900
- mirror: http://www.bilibili.com/video/av16594836/
Improved Explainability of Capsule Networks: Relevance Path by Agreement
- intro: Concordia University & University of Toronto
- arxiv: https://arxiv.org/abs/1802.10204
Low Light
Exploring Image Enhancement for Salient Object Detection in Low Light Images
- intro: ACM Transactions on Multimedia Computing, Communications, and Applications
- arxiv: https://arxiv.org/abs/2007.16124
NOD: Taking a Closer Look at Detection under Extreme Low-Light Conditions with Night Object Detection Dataset
- intro: BMVC 2021
- arxiv: https://arxiv.org/abs/2110.10364
Computer Vision
A Taxonomy of Deep Convolutional Neural Nets for Computer Vision
On the usability of deep networks for object-based image analysis
- intro: GEOBIA 2016
- arxiv: http://arxiv.org/abs/1609.06845
Learning Recursive Filters for Low-Level Vision via a Hybrid Neural Network
- intro: ECCV 2016
- project page: http://www.sifeiliu.net/linear-rnn
- paper: http://faculty.ucmerced.edu/mhyang/papers/eccv16_rnn_filter.pdf
- poster: http://www.eccv2016.org/files/posters/O-3A-03.pdf
- github: https://github.com/Liusifei/caffe-lowlevel
Toward Geometric Deep SLAM
- intro: Magic Leap, Inc
- arxiv: https://arxiv.org/abs/1707.07410
Learning Dual Convolutional Neural Networks for Low-Level Vision
- intro: CVPR 2018
- arxiv: https://arxiv.org/abs/1805.05020
Not just a matter of semantics: the relationship between visual similarity and semantic similarity
https://arxiv.org/abs/1811.07120
DF-SLAM: A Deep-Learning Enhanced Visual SLAM System based on Deep Local Features
- intro: BUPT & Megvii
- arxiv: https://arxiv.org/abs/1901.07223
GN-Net: The Gauss-Newton Loss for Deep Direct SLAM
- intro: Technical University of Munich & Artisense
- arxiv: https://arxiv.org/abs/1904.11932
All-In-One Network
HyperFace: A Deep Multi-task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition
- arxiv: https://arxiv.org/abs/1603.01249
- summary: https://github.com/aleju/papers/blob/master/neural-nets/HyperFace.md
UberNet: Training a `Universal’ Convolutional Neural Network for Low-, Mid-, and High-Level Vision using Diverse Datasets and Limited Memory
An All-In-One Convolutional Neural Network for Face Analysis
- intro: simultaneous face detection, face alignment, pose estimation, gender recognition, smile detection, age estimation and face recognition
- arxiv: https://arxiv.org/abs/1611.00851
MultiNet: Real-time Joint Semantic Reasoning for Autonomous Driving
- intro: first place on Kitti Road Segmentation. joint classification, detection and semantic segmentation via a unified architecture, less than 100 ms to perform all tasks
- arxiv: https://arxiv.org/abs/1612.07695
- github: https://github.com/MarvinTeichmann/MultiNet
Adversarial Collaboration: Joint Unsupervised Learning of Depth, Camera Motion, Optical Flow and Motion Segmentation
https://arxiv.org/abs/1805.09806
Visual Person Understanding through Multi-Task and Multi-Dataset Learning
- intro: RWTH Aachen University
- arxiv: https://arxiv.org/abs/1906.03019
Deep Learning for Data Structures
The Case for Learned Index Structures
- intro: MIT & Google
- keywords: B-Tree-Index, Hash-Index, BitMap-Index
- arxiv: https://arxiv.org/abs/1712.01208
Projects
Top Deep Learning Projects
deepnet: Implementation of some deep learning algorithms
DeepNeuralClassifier(Julia): Deep neural network using rectified linear units to classify hand written digits from the MNIST dataset
Clarifai Node.js Demo
- github: https://github.com/patcat/Clarifai-Node-Demo
- blog(“How to Make Your Web App Smarter with Image Recognition”): http://www.sitepoint.com/how-to-make-your-web-app-smarter-with-image-recognition/
Deep Learning in Rust
- blog(“baby steps”): https://medium.com/@tedsta/deep-learning-in-rust-7e228107cccc#.t0pskuwkm
- blog(“a walk in the park”): https://medium.com/@tedsta/deep-learning-in-rust-a-walk-in-the-park-fed6c87165ea#.pucj1l5yx
- github: https://github.com/tedsta/deeplearn-rs
Implementation of state-of-art models in Torch
Deep Learning (Python, C, C++, Java, Scala, Go)
deepmark: THE Deep Learning Benchmarks
Siamese Net
- intro: “This package shows how to train a siamese network using Lasagne and Theano and includes network definitions for state-of-the-art networks including: DeepID, DeepID2, Chopra et. al, and Hani et. al. We also include one pre-trained model using a custom convolutional network.”
- github: https://github.com/Kadenze/siamese_net
PRE-TRAINED CONVNETS AND OBJECT LOCALISATION IN KERAS
- blog: https://blog.heuritech.com/2016/04/26/pre-trained-convnets-and-object-localisation-in-keras/
- github: https://github.com/heuritech/convnets-keras
Deep Learning algorithms with TensorFlow: Ready to use implementations of various Deep Learning algorithms using TensorFlow
- homepage: http://www.gabrieleangeletti.com/
- github: https://github.com/blackecho/Deep-Learning-TensorFlow
Fast Multi-threaded VGG 19 Feature Extractor
Live demo of neural network classifying images
http://ml4a.github.io/dev/demos/cifar_confusion.html#
mojo cnn: c++ convolutional neural network
- intro: the fast and easy header only c++ convolutional neural network package
- github: https://github.com/gnawice/mojo-cnn
DeepHeart: Neural networks for monitoring cardiac data
Deep Water: Deep Learning in H2O using Native GPU Backends
- intro: Native implementation of Deep Learning models for GPU backends (mxnet, Caffe, TensorFlow, etc.)
- github: https://github.com/h2oai/deepwater
Greentea LibDNN: Greentea LibDNN - a universal convolution implementation supporting CUDA and OpenCL
Dracula: A spookily good Part of Speech Tagger optimized for Twitter
- intro: A deep, LSTM-based part of speech tagger and sentiment analyser using character embeddings instead of words. Compatible with Theano and TensorFlow. Optimized for Twitter.
- homepage: http://dracula.sentimentron.co.uk/
- speech tagging demo: http://dracula.sentimentron.co.uk/pos-demo/
- sentiment demo: http://dracula.sentimentron.co.uk/sentiment-demo/
- github: https://github.com/Sentimentron/Dracula
Trained image classification models for Keras
- intro: Keras code and weights files for popular deep learning models.
- intro: VGG16, VGG19, ResNet50, Inception v3
- github: https://github.com/fchollet/deep-learning-models
PyCNN: Cellular Neural Networks Image Processing Python Library
regl-cnn: Digit recognition with Convolutional Neural Networks in WebGL
- intro: TensorFlow, WebGL, regl
- github: https://github.com/Erkaman/regl-cnn/
- demo: https://erkaman.github.io/regl-cnn/src/demo.html
dagstudio: Directed Acyclic Graph Studio with Javascript D3
NEUGO: Neural Networks in Go
gvnn: Neural Network Library for Geometric Computer Vision
DeepForge: A development environment for deep learning
Implementation of recent Deep Learning papers
- intro: DenseNet / DeconvNet / DenseRecNet
- github: https://github.com/tdeboissiere/DeepLearningImplementations
GPU-accelerated Theano & Keras on Windows 10 native
Head Pose and Gaze Direction Estimation Using Convolutional Neural Networks
Intel(R) Math Kernel Library for Deep Neural Networks (Intel(R) MKL-DNN)
- homepage: https://01.org/mkl-dnn
- github: https://github.com/01org/mkl-dnn
Deep CNN and RNN - Deep convolution/recurrent neural network project with TensorFlow
Experimental implementation of novel neural network structures
- intro: binarynet / ternarynet / qrnn / vae / gcnn
- github: https://github.com/DingKe/nn_playground
WaterNet: A convolutional neural network that identifies water in satellite images
Kur: Descriptive Deep Learning
- github: https://github.com/deepgram/kur
- docs: http://kur.deepgram.com/
Development of JavaScript-based deep learning platform and application to distributed training
- intro: Workshop paper for ICLR2017
- arxiv: https://arxiv.org/abs/1702.01846
- github: https://github.com/mil-tokyo
NewralNet
- intro: A lightweight, easy to use and open source Java library for experimenting with feed-forward neural nets and deep learning.
- gitlab: https://gitlab.com/flimmerkiste/NewralNet
FeatherCNN
- intro: FeatherCNN is a high performance inference engine for convolutional neural networks
- github: https://github.com/Tencent/FeatherCNN
Readings and Questions
What you wanted to know about AI
http://fastml.com/what-you-wanted-to-know-about-ai/
Epoch vs iteration when training neural networks
- stackoverflow: http://stackoverflow.com/questions/4752626/epoch-vs-iteration-when-training-neural-networks
Questions to Ask When Applying Deep Learning
http://deeplearning4j.org/questions.html
How can I know if Deep Learning works better for a specific problem than SVM or random forest?
What is the difference between deep learning and usual machine learning?
Resources
Awesome Deep Learning
Awesome-deep-vision: A curated list of deep learning resources for computer vision
- website: http://jiwonkim.org/awesome-deep-vision/
- github: https://github.com/kjw0612/awesome-deep-vision
Applied Deep Learning Resources: A collection of research articles, blog posts, slides and code snippets about deep learning in applied settings.
Deep Learning Libraries by Language
Deep Learning Resources
http://yanirseroussi.com/deep-learning-resources/
Deep Learning Resources
https://omtcyfz.github.io/2016/08/29/Deep-Learning-Resources.html
Turing Machine: musings on theory & code(DEEP LEARNING REVOLUTION, summer 2015, state of the art & topnotch links)
BICV Group: Biologically Inspired Computer Vision research group
http://www.bicv.org/deep-learning/
Learning Deep Learning
http://rt.dgyblog.com/ref/ref-learning-deep-learning.html
Summaries and notes on Deep Learning research papers
Deep Learning Glossary
- intro: “Simple, opinionated explanations of various things encountered in Deep Learning / AI / ML.”
- author: Ryan Dahl, author of NodeJS.
- github: https://github.com/ry/deep_learning_glossary
The Deep Learning Playbook
https://medium.com/@jiefeng/deep-learning-playbook-c5ebe34f8a1a#.eg9cdz5ak
Deep Learning Study: Study of HeXA@UNIST in Preparation for Submission
Deep Learning Books
awesome-very-deep-learning: A curated list of papers and code about very deep neural networks (50+ layers)
Deep Learning Resources and Tutorials using Keras and Lasagne
Deep Learning: Definition, Resources, Comparison with Machine Learning
Awesome - Most Cited Deep Learning Papers
The most cited papers in computer vision and deep learning
deep learning papers: A place to collect papers that are related to deep learning and computational biology
papers-I-read
- intro: “I am trying a new initiative - a-paper-a-week. This repository will hold all those papers and related summaries and notes.”
- github: https://github.com/shagunsodhani/papers-I-read
LEARNING DEEP LEARNING - MY TOP-FIVE LIST
awesome-free-deep-learning-papers
DeepLearningBibliography: Bibliography for Publications about Deep Learning using GPU
- homepage: http://memkite.com/deep-learning-bibliography/
- github: https://github.com/memkite/DeepLearningBibliography
Deep Learning Papers Reading Roadmap
deep-learning-papers
- intro: Papers about deep learning ordered by task, date. Current state-of-the-art papers are labelled.
- github: https://github.com/sbrugman/deep-learning-papers/blob/master/README.md
Deep Learning and applications in Startups, CV, Text Mining, NLP
- github: https://github.com/lipiji/app-dl
ml4a-guides - a collection of practical resources for working with machine learning software, including code and tutorials
deep-learning-resources
- intro: A Collection of resources I have found useful on my journey finding my way through the world of Deep Learning.
- github: https://github.com/chasingbob/deep-learning-resources
21 Deep Learning Videos, Tutorials & Courses on Youtube from 2016
Awesome Deep learning papers and other resources
awesome-deep-vision-web-demo
- intro: A curated list of awesome deep vision web demo
- github: https://github.com/hwalsuklee/awesome-deep-vision-web-demo
Summaries of machine learning papers
https://github.com/aleju/papers
Awesome Deep Learning Resources
https://github.com/guillaume-chevalier/awesome-deep-learning-resources
Virginia Tech Vision and Learning Reading Group
https://github.com//vt-vl-lab/reading_group
MEGALODON: ML/DL Resources At One Place
- intro: Various ML/DL Resources organised at a single place.
- arxiv: https://github.com//vyraun/Megalodon
Arxiv Pages
Neural and Evolutionary Computing
https://arxiv.org/list/cs.NE/recent
Learning
https://arxiv.org/list/cs.LG/recent
Computer Vision and Pattern Recognition
https://arxiv.org/list/cs.CV/recent
Arxiv Sanity Preserver
- intro: Built by @karpathy to accelerate research.
- page: http://www.arxiv-sanity.com/
Today’s Deep Learning
http://todaysdeeplearning.com/
arXiv Analytics
Papers with Code
Papers with Code
Tools
DNNGraph - A deep neural network model generation DSL in Haskell
- homepage: http://ajtulloch.github.io/dnngraph/
Deep playground: an interactive visualization of neural networks, written in typescript using d3.js
- homepage: http://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle®Dataset=reg-plane&learningRate=0.03®ularizationRate=0&noise=0&networkShape=4,2&seed=0.23990&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification
- github: https://github.com/tensorflow/playground
Neural Network Package
- intro: This package provides an easy and modular way to build and train simple or complex neural networks using Torch
- github: https://github.com/torch/nn
deepdish: Deep learning and data science tools from the University of Chicago deepdish: Serving Up Chicago-Style Deep Learning
- homepage: http://deepdish.io/
- github: https://github.com/uchicago-cs/deepdish
AETROS CLI: Console application to manage deep neural network training in AETROS Trainer
- intro: Create, train and monitor deep neural networks using a model designer.
- homepage: http://aetros.com/
- github: https://github.com/aetros/aetros-cli
Deep Learning Studio: Cloud platform for designing Deep Learning AI without programming
cuda-on-cl: Build NVIDIA® CUDA™ code for OpenCL™ 1.2 devices
Receptive Field Calculator
- homepage: http://fomoro.com/tools/receptive-fields/
- example: http://fomoro.com/tools/receptive-fields/#3,1,1,VALID;3,1,1,VALID;3,1,1,VALID
receptivefield
- intro: (PyTorch/Keras/TensorFlow)Gradient based receptive field estimation for Convolutional Neural Networks
- github: https://github.com//fornaxai/receptivefield
Challenges / Hackathons
Open Images Challenge 2018
https://storage.googleapis.com/openimages/web/challenge.html
VisionHack 2017
- intro: 10 - 14 Sep 2017, Moscow, Russia
- intro: a full-fledged hackathon that will last three full days
- homepage: http://visionhack.misis.ru/
NVIDIA AI City Challenge Workshop at CVPR 2018
http://www.aicitychallenge.org/
Books
Deep Learning
- author: Ian Goodfellow, Aaron Courville and Yoshua Bengio
- homepage: http://www.deeplearningbook.org/
- website: http://goodfeli.github.io/dlbook/
- github: https://github.com/HFTrader/DeepLearningBook
- notes(“Deep Learning for Beginners”): http://randomekek.github.io/deep/deeplearning.html
Fundamentals of Deep Learning: Designing Next-Generation Artificial Intelligence Algorithms
- author: Nikhil Buduma
- book review: http://www.opengardensblog.futuretext.com/archives/2015/08/book-review-fundamentals-of-deep-learning-designing-next-generation-artificial-intelligence-algorithms-by-nikhil-buduma.html
- github: https://github.com/darksigma/Fundamentals-of-Deep-Learning-Book
FIRST CONTACT WITH TENSORFLOW: Get started with with Deep Learning programming
- author: Jordi Torres
- book: http://www.jorditorres.org/first-contact-with-tensorflow/
《解析卷积神经网络—深度学习实践手册》
- intro: by 魏秀参(Xiu-Shen WEI)
- homepage: http://lamda.nju.edu.cn/weixs/book/CNN_book.html
Make Your Own Neural Network: IPython Neural Networks on a Raspberry Pi Zero
- book: http://makeyourownneuralnetwork.blogspot.jp/2016/03/ipython-neural-networks-on-raspberry-pi.html
- github: https://github.com/makeyourownneuralnetwork/makeyourownneuralnetwork
Blogs
Neural Networks and Deep Learning
http://neuralnetworksanddeeplearning.com
Deep Learning Reading List
http://deeplearning.net/reading-list/
WILDML: A BLOG ABOUT MACHINE LEARNING, DEEP LEARNING AND NLP.
Andrej Karpathy blog
Rodrigob’s github page
colah’s blog
What My Deep Model Doesn’t Know…
http://mlg.eng.cam.ac.uk/yarin/blog_3d801aa532c1ce.html
Christoph Feichtenhofer
- intro: PhD Student, Graz University of Technology
- homepage: http://feichtenhofer.github.io/
Image recognition is not enough: As with language, photos need contextual intelligence
https://medium.com/@ken_getquik/image-recognition-is-not-enough-293cd7d58004#.dex817l2z
ResNets, HighwayNets, and DenseNets, Oh My!
- blog: https://medium.com/@awjuliani/resnets-highwaynets-and-densenets-oh-my-9bb15918ee32#.pgltg8pro
- github: https://github.com/awjuliani/TF-Tutorials/blob/master/Deep%20Network%20Comparison.ipynb
The Frontiers of Memory and Attention in Deep Learning
Design Patterns for Deep Learning Architectures
http://www.deeplearningpatterns.com/doku.php
Building a Deep Learning Powered GIF Search Engine
850k Images in 24 hours: Automating Deep Learning Dataset Creation
How six lines of code + SQL Server can bring Deep Learning to ANY App
- blog: https://blogs.technet.microsoft.com/dataplatforminsider/2017/01/05/how-six-lines-of-code-sql-server-can-bring-deep-learning-to-any-app/
- github: https://github.com/Microsoft/SQL-Server-R-Services-Samples/tree/master/Galaxies
Neural Network Architectures
- blog: https://medium.com/towards-data-science/neural-network-architectures-156e5bad51ba#.m8y39oih6
- blog: https://culurciello.github.io/tech/2016/06/04/nets.html