• Publications
  • Influence
Learning Structured Output Representation using Deep Conditional Generative Models
TLDR
In this work, we develop a deep conditional generative model for structured output prediction using Gaussian latent variables. Expand
  • 1,099
  • 270
  • PDF
Learning to Adapt Structured Output Space for Semantic Segmentation
TLDR
In this paper, we propose an adversarial learning method for domain adaptation in the context of semantic segmentation. Expand
  • 457
  • 121
  • PDF
Improved Deep Metric Learning with Multi-class N-pair Loss Objective
  • Kihyuk Sohn
  • Mathematics, Computer Science
  • NIPS
  • 5 December 2016
TLDR
We propose to address this problem with a new metric learning objective called multi-class N-pair loss. Expand
  • 644
  • 111
  • PDF
FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
TLDR
We introduce FixMatch, which produces artificial labels using both consistency regularization and pseudo-labeling and achieves state-of-the-art performance across a variety of standard semi-supervised learning benchmarks. Expand
  • 253
  • 71
  • PDF
Attribute2Image: Conditional Image Generation from Visual Attributes
TLDR
This paper investigates a novel problem of generating images from visual attributes. Expand
  • 551
  • 57
  • PDF
Understanding and Improving Convolutional Neural Networks via Concatenated Rectified Linear Units
TLDR
We propose a novel, simple yet effective activation scheme called concatenated ReLU (CRelu) and theoretically analyze its reconstruction property in CNNs. Expand
  • 297
  • 33
  • PDF
Augmenting CRFs with Boltzmann Machine Shape Priors for Image Labeling
TLDR
We propose the GLOC (GLObal and LOCal) model, a strong model for image labeling problems that combines the best properties of the CRF (that enforces local consistency between adjacent nodes) and the RBM (that models global shape prior of the 1 object). Expand
  • 155
  • 32
  • PDF
Towards Large-Pose Face Frontalization in the Wild
TLDR
We propose a novel deep 3D Morphable Model (3DMM) conditioned Face Frontalization Generative Adversarial Network (GAN), termed as FF-GAN, to generate neutral head pose face images. Expand
  • 199
  • 31
  • PDF
ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring
TLDR
We improve the recently-proposed "MixMatch" semi-supervised learning algorithm by introducing two new techniques: distribution alignment and augmentation anchoring. Expand
  • 102
  • 22
  • PDF
ReMixMatch: Semi-Supervised Learning with Distribution Matching and Augmentation Anchoring
TLDR
We improve the recently-proposed MixMatch semi-supervised learning algorithm by introducing two new techniques: distribution alignment and augmentation anchoring. Expand
  • 65
  • 22
...
1
2
3
4
5
...