收集主要整理了2016-2017年深度學習NLP、影象及語音等領域應用最前沿、最經典的論文。

目錄：

1 程式碼方面

1.1 程式碼生成

1.2 惡意軟體檢測/安全

2 NLP領域

2.1 摘要生成

2.2 Taskbots

2.3 分類

2.4 問答系統

2.5 情感分析

2.6 機器翻譯

2.7 聊天機器人

2.8 推理

3 計算及視覺方面

3.1 遊戲應用

3.2 風格遷移

3.3 物體跟蹤

3.4 視覺問答

3.5 影象分割

3.6 文字識別

3.7 大腦計算機互動

3.8 自動駕駛

3.9 物件識別

3.10 標識識別

3.11 高清晰度處理

3.12 姿態估計

3.13 影象標記

3.14 影象壓縮

3.15 影象合成

3.16 人臉識別

4 語音方面

4.1 語音合成

5 其他

5.1 未分類

5.2 正則

5.3 神經網路壓縮

5.4 優化器

1 程式碼方面

1.1 程式碼生成

DeepAM: Migrate APIs with Multi-modal Sequence to Sequence Learning 25 apr 2017

A Syntactic Neural Model for General-Purpose Code Generation 6 apr 2017

RobustFill: Neural Program Learning under Noisy I/O 21 mar 2017

DeepFix: Fixing Common C Language Errors by Deep Learning 12 feb 2017

DeepCoder: Learning to Write Programs 7 nov 2016

Neuro-Symbolic Program Synthesis 6 nov 2016

Deep API Learning 27 may 2016

1.2 惡意軟體檢測/安全

PassGAN: A Deep Learning Approach for Password Guessing 1 sep 2017

Deep Android Malware Detection 22 mar 2016 paper

Droid-Sec: Deep Learning in Android Malware Detection 17 aug 2014 paper

2 NLP領域

2.1 摘要生成

A Deep Reinforced Model for Abstractive Summarization 11 may 2017

Get To The Point: Summarization with Pointer-Generator Networks 14 apr 2017

SummaRuNNer: A Recurrent Neural Network based Sequence Model for Extractive Summarization of Documents 14 nov 2016

2.2 Taskbots

Composite Task-Completion Dialogue Policy Learning via Hierarchical Deep Reinforcement Learning 10 apr 2017 arxiv

End-to-End Task-Completion Neural Dialogue Systems 3 mar 2017 arxiv

2.3 分類

A Large Self-Annotated Corpus for Sarcasm 19 apr 2017

ConceptNet at SemEval-2017 Task 2: Extending Word Embeddings with Multilingual Relational Knowledge 11 apr 2017

Bilateral Multi-Perspective Matching for Natural Language Sentences 13 feb 2017

FastText.zip: Compressing text classification models 12 dec 2016

ConceptNet 5.5: An Open Multilingual Graph of General Knowledge 12 dec 2016

A Simple but Tough-to-Beat Baseline for Sentence Embeddings 4 nov 2016 paper

Enriching Word Vectors with Subword Information 15 jul 2016

From Word Embeddings To Document Distances 6 jul 2016 paper

Bag of Tricks for Efficient Text Classification 6 jul 2016

Character-level Convolutional Networks for Text Classification 4 sep 2015

GloVe: Global Vectors for Word Representation 25 may 2015 paper

Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks 28 feb 2015

Distributed Representations of Sentences and Documents 16 may 2014

Efficient Estimation of Word Representations in Vector Space 16 jan 2013

SimHash: Hash-based Similarity Detection 13 dec 2007

2.4 問答系統

IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models 30 may 2017 arxiv github state-of-the-art

2.5 情感分析

Rationalizing Neural Predictions 13 jun 2016

Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank 18 okt 2013

2.6 機器翻譯

Attention Is All You Need 12 jun 2017 arxiv state-of-the-art

Convolutional Sequence to Sequence Learning 8 may 2017 arxiv github state-of-the-art

Google's Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation 14 nov 2016

A Convolutional Encoder Model for Neural Machine Translation 7 nov 2016

Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation 26 sep 2016

Neural Machine Translation by Jointly Learning to Align and Translate 1 sep 2014

2.7 聊天機器人

A Deep Reinforcement Learning Chatbot 7 sep 2017

A Neural Conversational Model 19 jun 2015 arxiv

2.8 推理

Tracking the World State with Recurrent Entity Networks 12 dec 2016

3 計算及視覺方面

3.1 遊戲應用

Phase-Functioned Neural Networks for Character Control 1 may 2017

Equivalence Between Policy Gradients and Soft Q-Learning 21 apr 2017 arxiv deep-q-learning Beating Atari with Natural Language Guided Reinforcement Learning 18 apr 2017 arxiv deep-q-learning

Learning from Demonstrations for Real World Reinforcement Learning 12 apr 2017 arxiv deep-q-learning

FeUdal Networks for Hierarchical Reinforcement Learning 3 mar 2017 arxiv deep-q-learning

Overcoming catastrophic forgetting in neural networks 2 dec 2016 arxiv deep-q-learning

Playing Doom with SLAM-Augmented Deep Reinforcement Learning 1 dec 2016

Playing FPS Games with Deep Reinforcement Learning 18 sep 2016

DeepChess: End-to-End Deep Neural Network for Automatic Learning in Chess 16 aug 2016

Generative Adversarial Imitation Learning 10 jun 2016

Dueling Network Architectures for Deep Reinforcement Learning 20 nov 2015 arxiv deep-q-learning

Prioritized Experience Replay 18 nov 2015

Human-level control through deep reinforcement learning 26 feb 2015 paper deep-q-learning

Playing Atari with Deep Reinforcement Learning 19 dec 2013 arxiv deep-q-learning

3.2 風格遷移

Deep Photo Style Transfer 22 mar 2017

Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization 20 mar 2017 arxiv

A Learned Representation For Artistic Style 24 okt 2016

Instance Normalization: The Missing Ingredient for Fast Stylization 27 jul 2016

Perceptual Losses for Real-Time Style Transfer and Super-Resolution 27 mar 2016 arxiv

A Neural Algorithm of Artistic Style 26 aug 2015 arxiv

3.3 物體跟蹤

End-to-end representation learning for Correlation Filter based tracking 20 apr 2017 arxiv

3.4 視覺問答

VQA: Visual Question Answering 3 may 2015

3.5 影象分割

Dilated Residual Networks 22 jul 2017

SfM-Net: Learning of Structure and Motion from Video 25 apr 2017

Semi and Weakly Supervised Semantic Segmentation Using Generative Adversarial Network 28 mar 2017

Mask R-CNN 20 mar 2017 arxiv state-of-the-art

Learning Features by Watching Objects Move 19 dec 2016 arxiv state-of-the-art

RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation 20 nov 2016 arxiv

UberNet: Training a `Universal' Convolutional Neural Network for Low-, Mid-, and High-Level Vision using Diverse Datasets and Limited Memory 7 sep 2016

DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs 2 jun 2016

Fully Convolutional Networks for Semantic Segmentation 20 may 2016 arxiv

Instance-aware Semantic Segmentation via Multi-task Network Cascades 14 dec 2015

Multi-Scale Context Aggregation by Dilated Convolutions 23 nov 2015

SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation 2 nov 2015

U-Net: Convolutional Networks for Biomedical Image Segmentation 18 may 2015

Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs 22 dec 2014

Learning Rich Features from RGB-D Images for Object Detection and Segmentation 22 jul 2014

3.6 文字識別

OCR Error Correction Using Character Correction and Feature-Based Word Classification 21 apr 2016

Recursive Recurrent Nets with Attention Modeling for OCR in the Wild 9 mar 2016

COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images 26 jan 2016

Efficient Scene Text Localization and Recognition with Local Character Refinement 14 apr 2015

Reading Text in the Wild with Convolutional Neural Networks 4 dec 2014

Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition 9 jun 2014

3.7 大腦計算機互動

Deep learning with convolutional neural networks for brain mapping and decoding of movement-related information from the human EEG 15 mar 2017

Encoding Voxels with Deep Learning 2 dec 2015

Deep Neural Networks Reveal a Gradient in the Complexity of Neural Representations across the Ventral Stream 8 jul 2015

3.8 自動駕駛

Computer Vision for Autonomous Vehicles: Problems, Datasets and State-of-the-Art 18 apr 2017

End to End Learning for Self-Driving Cars 25 apr 2016

3.9 物體識別

Focal Loss for Dense Object Detection 7 aug 2017

Introspective Classifier Learning: Empower Generatively 25 apr 2017

Learning Chained Deep Features and Classifiers for Cascade in Object Detection 23 feb 2017

DSSD : Deconvolutional Single Shot Detector 23 jan 2017 arxiv state-of-the-art region-proposals

YOLO9000: Better, Faster, Stronger 25 dec 2016 arxiv github state-of-the-art region-proposals

Feature Pyramid Networks for Object Detection 9 dec 2016 arxiv state-of-the-art

Speed/accuracy trade-offs for modern convolutional object detectors 30 nov 2016

Aggregated Residual Transformations for Deep Neural Networks 16 nov 2016

Aggregated Residual Transformations for Deep Neural Networks 16 nov 2016

Hierarchical Object Detection with Deep Reinforcement Learning 11 nov 2016 arxiv deep-q-learning

Xception: Deep Learning with Depthwise Separable Convolutions 7 okt 2016

Learning to Make Better Mistakes: Semantics-aware Visual Food Recognition 1 okt 2016

Densely Connected Convolutional Networks 25 aug 2016

Residual Networks of Residual Networks: Multilevel Residual Networks 9 aug 2016

Context Matters: Refining Object Detection in Video with Recurrent Neural Networks 15 jul 2016

R-FCN: Object Detection via Region-based Fully Convolutional Networks 20 may 2016

Training Region-based Object Detectors with Online Hard Example Mining 12 apr 2016

T-CNN: Tubelets with Convolutional Neural Networks for Object Detection from Videos 9 apr 2016

Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning 23 feb 2016

Deep Residual Learning for Image Recognition 10 dec 2015

SSD: Single Shot MultiBox Detector 8 dec 2015 arxiv region-proposals

Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) 23 nov 2015

ParseNet: Looking Wider to See Better 15 jun 2015

You Only Look Once: Unified, Real-Time Object Detection 8 jun 2015 arxiv region-proposals

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks 4 jun 2015 arxiv region-proposals

Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification 6 feb 2015

Deep Image: Scaling up Image Recognition 13 jan 2015

Rich feature hierarchies for accurate object detection and semantic segmentation 11 nov 2013

Selective Search for Object Recognition 11 mar 2013

ImageNet Classification with Deep Convolutional Neural Networks 3 dec 2012 paper getting-started

3.10 標識識別

Deep Learning Logo Detection with Data Expansion by Synthesising Context 29 dec 2016

Automatic Graphic Logo Detection via Fast Region-based Convolutional Networks 20 apr 2016

LOGO-Net: Large-scale Deep Logo Detection and Brand Recognition with Deep Region-based Convolutional Networks 8 nov 2015

DeepLogo: Hitting Logo Recognition with the Deep Neural Network Hammer 7 okt 2015

3.11 高清晰度處理

Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network 16 sep 2016

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network 15 sep 2016

RAISR: Rapid and Accurate Image Super Resolution 3 jun 2016

Perceptual Losses for Real-Time Style Transfer and Super-Resolution 27 mar 2016 arxiv

Image Super-Resolution Using Deep Convolutional Networks 31 dec 2014

3.12 姿態估計

Forecasting Human Dynamics from Static Images 11 apr 2017

Fast Single Shot Detection and Pose Estimation 19 sep 2016

Convolutional Pose Machines 30 jan 2016

Flowing ConvNets for Human Pose Estimation in Videos 9 jun 2015

3.13 影象標記

Actor-Critic Sequence Training for Image Captioning 29 jun 2017

Detecting and Recognizing Human-Object Interactions 24 apr 2017

Deep Reinforcement Learning-based Image Captioning with Embedding Reward 12 apr 2017 arxiv deep-q-learning

Towards Diverse and Natural Image Descriptions via a Conditional GAN 17 mar 2017

Temporal Tessellation: A Unified Approach for Video Analysis 21 dec 2016 arxiv

Self-critical Sequence Training for Image Captioning 2 dec 2016

Generation and Comprehension of Unambiguous Object Descriptions 7 nov 2015

Show, Attend and Tell: Neural Image Caption Generation with Visual Attention 10 feb 2015

Long-term Recurrent Convolutional Networks for Visual Recognition and Description 17 nov 2014

3.14 影象壓縮

Full Resolution Image Compression with Recurrent Neural Networks 18 aug 2016

3.15 影象合成

A Neural Representation of Sketch Drawings 11 apr 2017

BEGAN: Boundary Equilibrium Generative Adversarial Networks 31 mar 2017 arxiv

Improved Training of Wasserstein GANs 31 mar 2017

Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks 30 mar 2017 arxiv

Wasserstein GAN 26 jan 2017

RenderGAN: Generating Realistic Labeled Data 4 nov 2016

Conditional Image Generation with PixelCNN Decoders 16 jun 2016

Pixel Recurrent Neural Networks 25 jan 2016

Generative Adversarial Networks 10 jun 2014

3.16 人臉識別

Accessorize to a Crime: Real and Stealthy Attacks on State-of-the-Art Face Recognition 24 okt 2016

OpenFace: A general-purpose face recognition library with mobile applications 1 jun 2016

Emotion Recognition in the Wild via Convolutional Neural Networks and Mapped Binary Patterns 9 nov 2015

Deep Face Recognition 7 sep 2015

Compact Convolutional Neural Network Cascade for Face Detection 6 aug 2015

Learning Robust Deep Face Representation 17 jul 2015

Facenet: A unified embedding for face recognition and clustering 12 jun 2015

Multi-view Face Detection Using Deep Convolutional Neural Networks 10 feb 2015

4 語音方面

4.1 語音合成

Title Date Paper Code Labels

Deep Cross-Modal Audio-Visual Generation 26 apr 2017

A Neural Parametric Singing Synthesizer 12 apr 2017

Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders 5 apr 2017 arxiv

Tacotron: Towards End-to-End Speech Synthesis 29 mar 2017 arxiv

Deep Voice: Real-time Neural Text-to-Speech 25 feb 2017

WaveNet: A Generative Model for Raw Audio 12 sep 2016 arxiv

5 其他

5.1 未分類

A simple neural network module for relational reasoning 5 jun 2017 arxiv state-of-the-art

Deep Complex Networks 27 may 2017 arxiv

Learning to Fly by Crashing 19 apr 2017

Who Said What: Modeling Individual Labelers Improves Classification 26 mar 2017

Semi-supervised Knowledge Transfer for Deep Learning from Private Training Data 18 okt 2016

DeepMath - Deep Sequence Models for Premise Selection 14 jun 2016

Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue 16 mar 2016

Long Short-Term Memory 15 nov 1997

5.2 正則

Self-Normalizing Neural Networks 8 jun 2017 arxiv state-of-the-art

Concrete Dropout 22 may 2017 arxiv

Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning 6 jun 2015

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift 11 feb 2015

5.3 神經網路壓縮

Design of Efficient Convolutional Layers using Single Intra-channel Convolution, Topological Subdivisioning and Spatial "Bottleneck" Structure 15 aug 2016

SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size 24 feb 2016

Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding 1 okt 2015

5.4 優化器

Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour 8 jun 2017 arxiv state-of-the-art

Equilibrated adaptive learning rates for non-convex optimization 15 feb 2015

Adam: A Method for Stochastic Optimization 22 dec 2014

Deep learning with Elastic Averaging SGD 20 dec 2014

ADADELTA: An Adaptive Learning Rate Method 22 dec 2012

Advances in Optimizing Recurrent Networks 4 dec 2012

Efficient Backprop

更多深度學習在NLP方面應用的經典論文、實踐經驗和最新訊息，歡迎關注微信公眾號“深度學習與NLP”或“DeepLearning_NLP”或掃描二維碼新增關注。

往期精品內容推薦：