Trending Research

Generative Modeling by Estimating Gradients of the Data Distribution

12 Jul 2019ermongroup/ncsn

We introduce a new generative model where samples are produced via Langevin dynamics using gradients of the data distribution estimated with score matching.

IMAGE GENERATION IMAGE INPAINTING

74
1.37 stars / hour

Sliced Score Matching: A Scalable Approach to Density and Score Estimation

17 May 2019ermongroup/ncsn

However, it has been so far limited to simple, shallow models or low-dimensional data, due to the difficulty of computing the Hessian of log-density functions.

74
1.37 stars / hour

Med3D: Transfer Learning for 3D Medical Image Analysis

1 Apr 2019Tencent/MedicalNet

The performance on deep learning is significantly affected by volume of training data.

3D MEDICAL IMAGING SEGMENTATION TRANSFER LEARNING

58
1.11 stars / hour

Realistic-Neural-Talking-Head-Models

20 May 2019vincent-thevenin/Realistic-Neural-Talking-Head-Models

My implementation of Few-Shot Adversarial Learning of Realistic Neural Talking Head Models (Egor Zakharov et al.).

META-LEARNING ONE-SHOT LEARNING TALKING HEAD GENERATION

121
0.93 stars / hour

R-Transformer: Recurrent Neural Network Enhanced Transformer

12 Jul 2019DSE-MSU/R-transformer

Recurrent Neural Networks have long been the dominating choice for sequence modeling.

LANGUAGE MODELLING MUSIC MODELING SEQUENTIAL IMAGE CLASSIFICATION

118
0.90 stars / hour

A Baseline for 3D Multi-Object Tracking

9 Jul 2019xinshuoweng/AB3DMOT

Although our baseline system is a straightforward combination of standard methods, we obtain the state-of-the-art results.

3D MULTI-OBJECT TRACKING AUTONOMOUS DRIVING

274
0.81 stars / hour
409
0.72 stars / hour

Unsupervised Data Augmentation

arXiv 2019 google-research/uda

Unlike previous methods that use random noise such as Gaussian noise or dropout noise, UDA has a small twist in that it makes use of harder and more realistic noise generated by state-of-the-art data augmentation methods.

DATA AUGMENTATION SENTIMENT ANALYSIS TEXT CLASSIFICATION

409
0.72 stars / hour

Gated-SCNN: Gated Shape CNNs for Semantic Segmentation

12 Jul 2019nv-tlabs/GSCNN

Here, we propose a new two-stream CNN architecture for semantic segmentation that explicitly wires shape information as a separate processing branch, i. e. shape stream, that processes information in parallel to the classical stream.

SEMANTIC SEGMENTATION

53
0.71 stars / hour