Powered by

# deep-learning-papers

# Computer Vision

# CNN Architecture

  • AlexNet: ImageNet Classification with Deep Convolutional Neural Networks (note)
  • ZFNet (DeconvNet): Visualizing and Understanding Convolutional Networks (note, code)
  • NIN: Network in Network
  • VggNet: Very Deep Convolutional Networks for Large-Scale Image Recognition
  • GoogLeNet: Going Deeper with Convolutions
  • ResNet:
    • ResNet-v1: Deep Residual Learning for Image Recognition (note)
    • ResNet-v2: Identity Mappings in Deep Residual Networks
    • Wide Residual Networks (note, code)
  • InceptionNet:
    • Inception-v1: GoogLeNet
    • Inception-v2, v3: Rethinking the Inception Architecture for Computer Vision (note, code)
    • Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
  • DenseNet:
  • NASNet: Learning Transferable Architectures for Scalable Image Recognition (note, code)
  • EfficientNet:(note)

# Visualizing CNNs

  • DeconvNet
  • BP: Deep inside convolutional networks: Visualising image classification models and saliency maps (note)
  • Guided-BP (DeconvNet+BP): Striving for simplicity: The all convolutional net (note, code)
  • Understanding Neural Networks Through Deep Visualization

# Weakly Supervised Localization

  • From Image-level to Pixel-level Labeling with Convolutional Networks (2015)
  • GMP-CAM: Is object localization for free? - Weakly-supervised learning with convolutional neural networks (2015) (note, code)
  • GAP-CAM: Learning Deep Features for Discriminative Localization (2016) (note, code)
  • c-MWP: Top-down Neural Attention by Excitation Backprop
  • Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization (2017) (note, code)

# Object Detection

  • OverFeat - Integrated Recognition, Localization and Detection using Convolutional Networks (note, code)

# Semantic Segmentation

  • FCN_V1 (2014)에서 직접적인 영향을 받은 모델들:
    • FCN + max-pooling indices를 사용: SegNet V2 (2015) (note)
    • FCN 개선: Fully Convolutional Networks for Semantic Segmentation (FCN_V2, 2016) (note, code)
    • FCN + atrous convolution과 CRF를 사용: DeepLap V2 (2016)
    • FCN + Dilated convolutions 사용: Multi-Scale Context Aggregation by Dilated Convolutions (2015)
    • FCN + Multi-scale: Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-Scale Convolutional Architecture (2015)
    • FCN + global context 반영 위해 global pooling 사용: ParseNet (2015)
    • FCN + 모든 레이어에서 skip 사용: U-Net (2015) (note)
  • PSPNet (2016) (note)
  • DeepLabv3+ (2018) (note)
  • EncNet (2018)
  • FastFCN (2019)
  • Instance Segmentation
    • DeepMask
    • SharpMask
    • Mask R-CNN (2017) (note)
  • 3D / Point Cloud
    • PointNet (2017)
    • SGPN (2017)
  • Weakly-supervised Segmentation

# Style Transfer

  • A Neural Algorithm of Artistic Style (2015)
  • Image Style Transfer Using Convolutional Neural Networks (2016)
  • Perceptual Losses for Real-Time Style Transfer and Super-Resolution (2016)
  • Instance Normalization:
    • Instance Normalization: The Missing Ingredient for Fast Stylization (2016)
    • Improved Texture Networks: Maximizing Quality and Diversity in Feed-forward Stylization and Texture Synthesis (2017)

# Siamese, Triplet Network

  • Triplet Network
    • FaceNet: A Unified Embedding for Face Recognition and Clustering (note, code)
    • Learning Fine-grained Image Similarity with Deep Ranking (note, code)
  • Siamese Network

# Mobile

  • Shufflenet: An extremely efficient convolutional neural network for mobile devices
  • Mobilenets: Efficient convolutional neural networks for mobile vision applications

# Etc.

  • A guide to convolution arithmetic for deep learning (note)

# Generative Models

# Models

# Autoregressive Models

# Variational Autoencoders

# Normalizing Flow Models

  • NICE (2014) (note)
  • Variational Inference with Normalizing Flows (2015) (code)
  • IAF (2016)
  • MAF (2017)
  • Glow (2018)

# GANs

  • GAN: Generative Adversarial Networks (note, code1(mnist), code2(fashion_mnist))
  • DCGAN (note, code1, code2)
  • WGAN 계열:
    • WGAN: Wasserstein GAN (note(진행중), code)
    • WGAN_GP: Improved Training of Wasserstein GANs
    • CT-GAN: Improving the Improved Training of Wasserstein GANs
  • infoGAN
  • Improved GAN:
  • SNGAN: Spectral Normalization for Generative Adversarial Networks (note(진행중), code)
  • SAGAN:
  • CoGAN: Coupled Generative Adversarial Networks (note, code)

# Image generation

# image-to-image

  • cGAN: Conditional Generative Adversarial Nets (2014) (note, code)
  • (내맘대로)pix2pix 계열:
    • pix2pix: Image-to-Image Translation with Conditional Adversarial Networks (2016) (note)
    • pix2pixHD
    • CycleGAN:
    • BicycleGAN
    • vid2vid: Video-to-Video Synthesis
    • SPADE: Semantic Image Synthesis with Spatially-Adaptive Normalization (2019) (note)
  • StarGAN:
  • PGGAN:
  • UNIT/MUNIT
  • iGAN
  • StyleGAN:

# text-to-image

  • Generative adversarial text to image synthesis
  • StackGAN
  • AttnGAN

# Sequence generation

  • WaveGAN: (note, code)
  • SeqGAN:

# Evaluation

  • A note on the evaluation of generative models
  • A Note on the Inception Score

# CS236 (Deep Generative Models)

  • Introduction and Background (slide 1, slide 2)
  • Autoregressive Models (slide 3, slide 4)
  • Variational Autoencoders
  • Normalizing Flow Models
  • Generative Adversarial Networks
  • Energy-based models

# NLP

  • Recent Trends in Deep Learning Based Natural Language Processing (note)

# RNN Architecture

  • Seq2Seq
    • Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation (2014)
    • Sequence to Sequence Learning with Neural Networks (2014) (note, code)
    • A Neural Conversational Model
  • Attention
    • (Luong) Effective Approaches to Attention-based Neural Machine Translation (2015)
    • (Bahdanau) Neural Machine Translation by Jointly Learning to Align and Translate (2014) (note, code)
    • Transformer: Attention Is All You Need (2017)
  • Memory Network
    • Memory Networks (2014)
    • End-To-End Memory Networks (2015)
  • Residual Connection
    • Deep Recurrent Models with Fast-Forward Connections for NeuralMachine Translation (2016)
    • Google's Neural MachineTranslation: Bridging the Gap between Human and Machine Translation (2016)
  • CNN
    • Convolutional Neural Networks for Sentence Classification (2014)
    • ByteNet: Neural Machine Translation in Linear Time (2017)
    • Depthwise Separable Convolutions for Neural Machine Translation (2017)
    • SliceNet: Convolutional Sequence to Sequence Learning (2017)

# Word Embedding

# Multimodal Learning

  • DeVise - A Deep Visual-Semantic Embedding Model: (note)
  • Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models: (note)
  • Show and Tell: (note)
  • Show, Attend and Tell: (note)
  • Multimodal Machine Learning: A Survey and Taxonomy: (note)

# Etc. (Optimization, Normalization, Applications)

  • An overview of gradient descent optimization algorithms (note)
  • Dropout:
  • Batch Normalization: (pdf+memo, code)
  • How Does Batch Normalization Help Optimization?
  • Spectral Norm Regularization for Improving the Generalizability of Deep Learning (note(진행중), code)
  • Wide & Deep Learning for Recommender Systems
  • Xavier Initialization - Understanding the difficulty of training deep feedforward neural networks
  • PReLU, He Initialization - Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification

# Drug Discovery

© Copyright 2020 - 2022 KyuHwan Shim, All rights reserved.