rock913 / paper-on-AI

This repository serves as a place to summarize the AI related papers I read

Paper-on-AI

Topics

Video Surveillance (Note)

Medical Image analysis using deep learning (Note)

Probabilistic programming language (Note)

• Human-level concept learning through probabilistic program induction (Note)
• Picture: A Probabilistic Programming Language for Scene Perception (Note)
• Deep API Programmer: Learning to Program with APIs We present DAPIP, a Programming-By-Example system that learns to program with APIs to perform data transformation tasks. We design a domain specific language (DSL) that allows for arbitrary concatenations of API outputs and constant strings. The DSL consists of three family of APIs: regular expression-based APIs, lookup APIs, and transformation APIs. We then present a novel neural synthesis algorithm to search for programs in the DSL that are consistent with a given set of examples. The search algorithm uses recently introduced neural architectures to encode input-output examples and to model the program search in the DSL. We show that synthesis algorithm outperforms baseline methods for synthesizing programs on both synthetic and real-world benchmarks.
  Neuro-Symbolic Program Synthesis

Bayesian Deep Learning

• ZhuSuan: A Library for Bayesian Deep Learning
• Robust RegBayes: Selectively Incorporating First-Order Logic Domain Knowledge into Bayesian Models
• Uncertainty in Deep Learning

Reinforcement Learning

• Becca: a general learning program for use in any robot or embodied system
• DEEP REINFORCEMENT LEARNING: AN OVERVIEW
• A Unified Game-Theoretic Approach to Multiagent Reinforcement Learning
• Deep Inverse Reinforcement Learning to learn the reward for each state
• Cooperative Inverse Reinforcement Learning

for finance

• async-rl
• A3C-Gym

Imbalanced Classification (Note)

• Learning Deep Representation for Imbalanced Classification_cvpr2016 (Note)
• Deep Over-sampling Framework for Classifying Imbalanced Data (Note)
• Metric Learning with Adaptive Density Discrimination
• Semi-supervised deep learning by metric embedding
• tf-magnet
• Training Neural Networks with Very Little Data(data augumentatin)

semi-supervised learning

• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results [code]
• Temporal Ensembling for Semi-Supervised Learning
• Virtual Adversarial Training: a Regularization Method for Supervised and Semi-supervised Learning
• Self-ensembling for domain adaptation
• Semi-supervised deep learning by metric embedding
• Semi-Supervised Learning with Generative Adversarial Networks
• Semi-Supervised Learning with Deep Generative Models Github1
• Semi-Supervised Learning with Deep Generative Models Github2
• Learning Loss Functions for Semi-supervised Learning via Discriminative Adversarial Networks
• Auxiliary Deep Generative Models

label noise

• CleanNet: Transfer Learning for Scalable Image Classifier Training with Label Noise

data ploting

• ImageTSNEViewer
• plot_embedding
• plot_embedding2

self play

• Mastering the Game of Go without Human Knowledge

vision

Hierarchical Surface Prediction for 3D Object Reconstruction 3D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstruction

face recognition

• Disentangled Representation Learning GAN for Pose-Invariant Face Recognition Github
• ArcFace: Additive Angular Margin Loss for Deep Face Recognition Github

health

Classification of Radiology Reports Using Neural Attention Models

Garment

• Physics-driven Pattern Adjustment for Direct 3D Garment Editing

• Detailed Garment Recovery from a Single-View Image
  Fig. 1. Garment recovery and re-purposing results. From left to right, we show an example of (a) the original image [Saaclothes 2015] c©, (b) the recovered dress and body shape from a single-view image, and (c)-(e) the recovered garment on another body of different poses and shapes/sizes [Hillsweddingdress 2015] c©.
  
  Fig. 2. The flowchart of our algorithm. We take a single-view image [ModCloth 2015] c©, a human-body dataset, and a garment-template database as input. We preprocess the input data by performing garment parsing, sizing and features estimation, and human-body reconstruction. Next, we recover an estimated garment described by the set of garment parameters, including fabric material, design pattern parameters, sizing and wrinkle density, as well as the registered garment dressed on the reconstructed body.
  

• Styling Evolution for Tight-Fitting Garments

• Parsing Sewing Patterns into 3D Garments

video game

• BUILDING GENERALIZABLE AGENTS WITH A REALISTIC AND RICH 3D ENVIRONMENT
• Artificial Intelligence and Games
• The Game Imitation: Deep Supervised Convolutional Networks for Quick Video Game AI
• ResearchDoom and CocoDoom: Learning Computer Vision with Games
• ELF: An Extensive, Lightweight and Flexible Research Platform for Real-time Strategy Games
• Beating the World’s Best at Super Smash Bros. Melee with Deep Reinforcement Learning
• A Deep Hierarchical Approach to Lifelong Learning in Minecraft
• General Video Game AI: Competition, Challenges, and Opportunities
• The General Video Game AI Competition - 2017
• Implementing Reinforcement Learning in Unreal Engine 4 with Blueprint
• Learning Policies for First Person Shooter Games Using Inverse Reinforcement Learning
• Strategy Detection in Wuzzit: A Decision Theoretic Approach

cloud spot instance Management

• Deep Learning on Amazon EC2 Spot Instances Without the Agonizing Pain
• portal-gun
• Learning Machine Learning on the cheap: Persistent AWS Spot Instances
• Persisting state between AWS EC2 spot instances

generate data from video game

• Playing for Data: Ground Truth from Computer Games
• UnrealCV: Connecting Computer Vision to Unreal Engine
• Augmented Reality Meets Computer Vision : Efficient Data Generation for Urban Driving Scenes
• Model-driven Simulations for Computer Vision
• UE4Sim: A Photo-Realistic Simulator for Computer Vision Applications
• Teaching UAVs to Race Using UE4Sim
• Video Propagation Networks

emotion recognition

• Emotion Recognition in the Wild using Deep Neural Networks and Bayesian Classifiers
• EmotioNet Challenge: Recognition of facial expressions of emotion in the wild

activity Recognition

• Learning to Detect Human-Object Interactions
• AVA: A Video Dataset of Spatio-temporally Localized Atomic Visual Actionschallenge
• The Kinetics Human Action Video Dataset

video tracking

• Deep Feature Flow for Video Recognition

• Flow-Guided Feature Aggregation for Video Object Detection

• Detect to Track and Track to Detect

• Simple Baselines for Human Pose Estimation and Trackinggithub

• Pose Flow: Efficient Online Pose Tracking

• PoseFlow

• Tracking The Untrackable:Learning to Track Multiple Cues with Long-Term Dependencies

• Moving beyond: Deepomatic learns how to track multiple objects

• signality

• ArtTrack: Articulated Multi-person Tracking in the Wild

background substraction fill

• awesome-background-subtraction
• FgSegNet
• FusionSeg
• generative_inpainting

reid

• DukeMTMC-VideoReID
• Clothing Change Aware Person Identification
• Beyond Frontal Faces: Improving Person Recognition Using Multiple Cues

persion re Identifying

• deep-person-reid
• AlignedReID: Surpassing Human-Level Performance in Person Re-Identification github
• Alignedreid＋+: Dynamically Matching Local Information for Person Re-Identification

scalable VideoReID

• Scalable Person Re-identification: A Benchmark
• Revisiting Temporal Modeling for Video-based Person ReID
• Exploiting Transitivity for Learning Person Re-identification Models on a Budget
• Deep Active Learning for Video-based Person Re-identification

person recognition

• A Multi-Level Contextual Model For Person Recognition in Photo Albums
• Large-scale Multimedia Analysis Project ReportPerson Recognition in Photo Albums
• Learning Deep Features via Congenerous Cosine Loss for Person Recognition
• github-Person Recognition System on PIPA dataset
• github
• Person Search in Videos with One Portrait Through Visual and Temporal Linkspage

code

• COCO Loss
• NormFace
• HumanRecognition
• Person recognition

Detection

• Flow-Guided Feature Aggregation for Video Object Detection
• Deep Feature Flow for Video Recognition
• SSD: Single Shot MultiBox Detector
• Spatially Supervised Recurrent Convolutional Neural Networks for Visual Object Tracking
• ROLO
• High-Speed Tracking-by-Detection Without Using Image Information
• End-to-end representation learning for Correlation Filter based tracking
• SIMPLE ONLINE AND REALTIME TRACKING
• ReMotENet: Efficient Relevant Motion Event Detection for Large-scale Home Surveillance Videos
• Multiple People Tracking by Lifted Multicut and Person Re-identification
• POI: Multiple Object Tracking with High Performance Detection and Appearance Featuregithub
• People Detection and Tracking in Crowded Scenes

annotation

• deepannotator
• annotation tool
• labelme
• An interactive tool for manual, semi-automatic and automatic video annotation
• Efficiently Scaling Up Video Annotation with Crowdsourced Marketplaces
• BeaverDam
• DeepTeach
• dash-object-detection
• action-annotation
• bbox-annotator
• Image Picker
• Django Photo Gallery Sample
• labelbox

annotation company

• [scale.ai]https://scale.ai/retail#how-it-works

datasets

• datasets

video health related

• Unconstrained Video Monitoring of Breathing Behavior and Application to Diagnosis of Sleep Apnea
• This Stanford-Tested Baby Monitor Uses Computer Vision To Keep Baby Safe
• cocooncam
• Vision-based patient monitoring: a comprehensive review of algorithms and technologies

GANS(Note)

• RENDERGAN: GENERATING REALISTIC LABELED DATA

conditional GAN

• Disentangled Variational Auto-Encoder for Semi-supervised Learning
• Fader Networks: Manipulating Images by Sliding Attributes Distangle face attributes from laten variables by GAN
• Real-valued (Medical) Time Series Generation with Recurrent Conditional GANs
• Learning to Discover Cross-Domain Relations with Generative Adversarial Networks
• FACE AGING WITH CONDITIONAL GENERATIVE ADVERSARIAL NETWORKS
• GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data
• Invertible Conditional GANs for image editing
• Image-to-Image Translation with Conditional Adversarial Networks

RNN

• Training RNNs as Fast as CNNs [code]
• CreativeAI

Reservoir Computing

• website
• A Comparative Study of Reservoir Computing for Temporal Signal Processing
• Samim

Design

-CreativeAI

###sports

• Data-Driven Ghosting using Deep Imitation Learning

text summorization

• Personal Research Agents on the Web of Linked Open Data
• GitXiv

web scraper

• web scraper blog1

• scrapper using GO

• A simple, higher level interface for Go web scraping

• Github scapper-python

• Github scapper-javascript

• How To Use node.js, request and cheerio to Set Up Simple Web-Scraping

abductive resoning

• Watson: Beyond Jeopardy! it elaborates upon a vision for an evidence-based clinical decision support system, based on the DeepQA technology, that affords exploration of a broad range of hypotheses and their associated evidence, as well as uncovers missing information that can be used in mixed-initiative dialog.
• Abduction in Machine Learning
• Logic Tensor Networks: Deep Learning and Logical Reasoning from Data and Knowledge
• Neural-Symbolic Computing, Deep logic networks and applications
• Reasoning with Deep Learning: an Open Challenge
• 3 Types Of Reasoning And AlphaGo inductive reasoning He died and she died. Everyone died, so I will die deductive reasoning If all humans die and I am a human, then I will die. abductive reasoning He died and the cat died, so he is a cat

Financial

• (paper list)[https://github.com/thuquant/awesome-quant/blob/master/papers.md)

• Deep Learning in Finance

• apis

• finviz

• Agent Inspired Trading Using Recurrent Reinforcement Learning and LSTM Neural Networks

• Deep Direct Reinforcement Learning for Financial Signal Representation and Trading

• MACHINE LEARNING FOR TRADING

• MOrdReD: Memory-based Ordinal Regression Deep Neural Networks for Time Series Forecasting

• Financial Time-Series Predictions and AI Models (Part 2): HTM Models

• Leveraging Financial News for Stock Trend Prediction with Attention-Based Recurrent Neural Network

portfolio management

• PGPortfolio
• A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem
• Deep Reinforcement Learning in Portfolio Management
• Gather data, compute statistics, and make predictions on securities
• etfpredict
• ETFDataLoader
• ETFGlobal
• Deep Reinforcement Learning in Portfolio ManagementGithub

combine news

• [Using Structured Events to Predict Stock Price Movement: An Empirical Investigation]9emnlp2014.org/papers/pdf/EMNLP2014148.pdf)
• Deep Learning for Event-Driven Stock Prediction
• On the Importance of Text Analysis for Stock Price Prediction
• Using Structured Events to Predict Stock Price Movement: An Empirical Investigation

Github

• financial-news-dataset

• PyDatastream

• Reuters-full-data-set

• 8k_stock_price_movement

• Predicting stock price directional movement after 8-K filings

• scraping sec.gov for Form 8-K's using Beautiful Soup

• Stock Market prediction using news headlines

• How To: Convert a Stock Ticker Into a CIK Identifier With Python

recursive network

• Recursive Neural Networks

EEG

• Generative Adversarial Networks Conditioned by Brain Signals
• Deep adversarial neural decoding

neuroscience

• NeuCube: A spiking neural network architecture for mapping, learning and understanding of spatio-temporal brain data

segmentation

• Speed/accuracy trade-offs for modern convolutional object detectors Good paper summory on instance segmentation methods:the Faster R-CNN, R-FCN and SSD systems
• Mask R-CNN The method, called Mask R-CNN, extends Faster R-CNN by adding a branch for predicting an object mask in parallel with the existing branch for bounding box recognition.
• The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation [code]

pose estimation

• Towards Accurate Multi-person Pose Estimation in the Wild
• Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks[code] [code2]
• Keep it SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Image
• Stacked Hourglass Networks for Human Pose Estimation cascade importance for pose estimation
• DensePose: Dense Human Pose Estimation In The Wild improved version of mask-rnn, good dataset for dense pose estimation

on Speed

• Pose Proposal Networks

fmri

• A Convolutional Autoencoder for Multi-Subject fMRI Data Aggregation a model using autoencoder, thinking about using deepsquezee for high accuracy

• A Reduced-Dimension fMRI Shared Response Model

• openfMRI(https://openfmri.org/how-to-extract-data/) large dataset

• Recurrent Neural Networks for Spatiotemporal Dynamics of Intrinsic Networks from fMRI Data

• Deep Learning with Edge Computing for Localization of Epileptogenicity using Multimodal rs-fMRI and EEG Big Data

###MRI registration

• An Unsupervised Learning Model for Deformable Medical Image Registration

medical imaing

• Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning

MRI Segmentation

• 3D fully convolutional networks for subcortical segmentation in MRI: A large-scale study
• HyperDense-Net: A densely connected CNN for multi-modal image segmentation
• Neuroimage special issue on brain segmentation and parcellation - Editorial
• End-to-end learning of brain tissue segmentation from imperfect labeling
• Scalable multimodal convolutional networks for brain tumour segmentation
• NiftyNet: a deep-learning platform for medical imaging
• DeepNAT: Deep Convolutional Neural Network for Segmenting Neuroanatomy

fmri-publicdata

• HCP
• DataLad

Capsules

• Dynamic Routing Between Capsules
• MATRIX CAPSULES WITH EM ROUTING

active learning

• libact: Pool-based Active Learning in Python (Github)
• Deep Bayesian Active Learning with Image Data
• Learning how to Active Learn: A Deep Reinforcement Learning Approach

autoencoder

• Neural Discrete Representation Learning

graph convolutional recurrent networks

• Structured sequence modeling with graph convolutional recurrent networks

one shot learning

• Matching Networks for One Shot Learning Some kind of similar with KNN, need to read a paper on "attention lstm"

Asynchronous Actor-Critic Agents

• Simple Reinforcement Learning with Tensorflow Part 8: Asynchronous Actor-Critic Agents (A3C)
• A major achievement in reinforcement learning research

database

• TensorLayer: A Versatile Library for Efficient Deep Learning Development
• tensorpack
• DeepSleepNet: a Model for Automatic Sleep Stage Scoring based on Raw Single-Channel EEG

automat learning process

• Learning to learn by gradient descent

Visualization

• Plug-and-Play Interactive Deep Network Visualization
• GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks
• NBA-player-movement

system examples

• home_surveillance
• A face recognition based attendance updating system
• A face recognition based attendance system

AI as servise (Note)

• labellio
• clarifai
• fastai
• deepdetect

Transfer and multi-task learning (Note)

transfer learning

• A survey of transfer learning
• Transfer learning and domain adaptationsecond frozone; fine-turn; domain; distillation; semi-supervised
• Transfer Learning — The Next Frontier for ML
• Transfer Learning - Machine Learning's Next Frontier

multi-task learning

• An Overview of Multi-Task Learning in Deep Neural Networks
• Multi-task Self-Supervised Visual Learning
• HyperFace: A Deep Multi-task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition
• Multi-task, Multi-lingual Learning

network compression

• MicroExpNet: An Extremely Small and Fast Model For Expression Recognition From Frontal Face Images
• SQUEEZENET: ALEXNET-LEVEL ACCURACY WITH 50X FEWER PARAMETERS AND <0.5MB MODEL SIZE

instresting website

• quid

blockchain

• atn
• ethereum
• AMCHART To Launch ICO On March 1 For Electronic Health Records (EHR) System[whitePaper]
  

vicarious

• vicarious
• Teaching Compositionality to CNNs
• Schema Networks: Zero-shot Transfer with a Generative Causal Model of Intuitive Physics
• A generative vision model that trains with high data efficiency and breaks text-based CAPTCHAs

academy research

• Microsoft Academic Graph
• An Overview of Microsoft Academic Service (MAS) and Applications

keywords analysis

• Towards robust tags for scientific publications from natural language processing tools and Wikipedia
• Semantic Tagging Using Topic Models Exploiting Wikipedia Category Network
• A 'supervised' parser for Google Scholar
• Meta-Scholar
• A Python module for extracting relevant tags from text documents.

public data

• Awesome Public Datasets

recommander system

• Deep Learning based Recommender System: A Survey and New Perspectives
• Wide & Deep Learning for Recommender Systems