Papers

Memos for papers, which are related to ML, CV and NLP.

CV

Recognition

• Wide Residual Networks
• Densely Connected Convolutional Networks
• Deep Pyramidal Residual Networks with Separated Stochastic Depth
• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• Dual Path Networks
• CondenseNet: An Efficient DenseNet using Learned Group Convolutions
• Recurrent Models of Visual Attention

Detection (Including Instance Segmentation)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• SSD: Single Shot MultiBox Detector
• Feature Pyramid Networks for Object Detection
• DSSD : Deconvolutional Single Shot Detector
• Speed/accuracy trade-offs for modern convolutional object detectors
• Focal Loss for Dense Object Detection
• DetNet: A Backbone network for Object Detection
• Light-Head R-CNN: In Defense of Two-Stage Object Detector
• Fully Convolutional Instance-aware Semantic Segmentation
• Mask R-CNN
• Fast and accurate object detection in high resolution 4K and 8K video using GPUs
• Revisiting RCNN: On Awakening the Classification Power of Faster RCNN

Pedestrian detection

• Faster R-CNN with Densenet for scale aware pedestrian detection vis-a-vis head negative suppression

Semantic Segmentation

• Fully Convolutional Networks for Semantic Segmentation
• SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation
• U-Net: Convolutional Networks for Biomedical Image Segmentation

Captioning

• Self-critical Sequence Training for Image Captioning
• Show and Tell: A Neural Image Caption Generator
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Deep visual-semantic alignments for generating image descriptions

GAN

• Generative Adversarial Nets
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• Image-to-Image Translation with Conditional Adversarial Networks
• cGAN-based Manga Colorization Using a Single Training Image
• Learning from Simulated and Unsupervised Images through Adversarial Training

Robust Reading

• TextBoxes++: A Single-Shot Oriented Scene Text Detector
• Synthetic data generation for Indic handwritten text recognition
• Reading Scene Text with Attention Convolutional Sequence Modeling

Visualization

• SmoothGrad: removing noise by adding noise

Video

Tracking

• Improving Online Multiple Object tracking with Deep Metric Learning
• SIMPLE ONLINE AND REALTIME TRACKING

Detection

• Mobile Video Object Detection with Temporally-Aware Feature Maps
• Towards High Performance Video Object Detection for Mobiles

Else

• Multiple Frames Matching for Object Discovery in Video

• Unsupervised Learning of Video Representations using LSTMs

• Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?

• DeepMark: One-Shot Clothing Detection

3D

• PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space

Else

• FlowNet: Learning Optical Flow with Convolutional Networks
• FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks
• Revisiting Unreasonable Effectiveness of Data in Deep Learning Era
• Learning to Compose Domain-Specific Transformations for Data Augmentation
• Spatial Transformer Networks

NLP

NMT

• Effective Approaches to Attention-based Neural Machine Translation
• Neural Machine Translation by Jointly Learning to Align and Translate
• Sequence to Sequence Learning with Neural Networks
• Attention Is All You Need

ML

• Positive-Unlabeled Learning with Non-Negative Risk Estimator

ELSE

• Unsupervised Deep Embedding for Clustering Analysis
• Attention-Based Models for Speech Recognition
• Bridging the Gaps Between Residual Learning, Recurrent Neural Networks and Visual Cortex
• What’s your ML Test Score? A rubric for ML production systems
• Multimodal Emoji Prediction
• Born Again Neural Networks
• Digital Auditor: A Framework for Matching Duplicate Invoices
• Pedestrian Detection: A Benchmark