Boqing Gong

Presented at CVPR 2020 Workshops on Learning from Imperfect Data and Adversarial Machine Learning in Computer Vision, respectively

Presented at CVPR 2020 Area Chair Workshop and WACV 2020 Workshop on Vision Applications and Solutions to Biased or Scarce Data

"Curriculum Domain Adaptation: Using Simulation for Real".

"The Multiple Shades of Dropout for Discriminative and Generative Deep Neural Networks". Dropout, which independently zeros out the outputs of neurons at random, has become one of the most popular techniques in training deep neural networks due to its simplicity and remarkable effectiveness. This talk reveals multiple shades of dropout for both discriminative and generative deep neural networks, mainly covering our following works: [Li et al., NIPS'16] and [Wei, Gong, et al., ICLR'18].

"Domain Adaptation and Transfer: All You Need to Use Simulation 'for Real'". Domain adaptation from simulation to the real world, and vice versa, for semantic segmentation and learning policies.

"Learning and Adapting from the Web for Visual Recognition".

"Learning from Web Data and Adapting Beyond It", which inlcudes the Web data of 1) noisy labels, 2) accurate labels, and 3) multi-modalities by semi-supervised learning, curriculum learning, and kernel mean matching, respectively. It mainly covers our following works: [Ding et al., WACV'18], [Wei, Gong, et al., ICLR'18], [Gan et al., CVPR'18], [Zhang et al., ICCV'17], [Gan et al., ECCV'16], and [Sharghi et al., ECCV'16].

Two at CVPR, two at ICCV, two book chapters, and one at ACII.

"Domain Adaptation for Robust Visual Recognition". It is about our following works: [Gong, et al., ICML'13], [Gan et al., ECCV'16], [Gan et al., CVPR'16], and [Zhang et al., ICCV'17].

"BIGDATA: IA: Distributed Semi-Supervised Training of Deep Models and Its Applications in Video Understanding" (transferred to Mubarak Shah and Liqiang Wang). This project investigates semi-supervised training of deep neural network models using large-scale labeled and unlabeled data in a distributed fashion.

for our researcho on "Face Detector Adaptation without Forgetting".

Three at IEEE CVPR, three at ECCV, one at NIPS, and one at Symposium Multemedia.

Project: "Collaborative Research: Florida-IT-Pathways to Success (Flit-Path)".

for our researcho on "User-Guided Visual Analytics".

"Domain Adaptation Approaches to Human Activity Detection, Recognition, and Summarization".

"CRII: RI: Multi-Source Domain Generalization Approaches to Visual Attribute Detection". This project investigates how to accurately and robustly detect attributes from images (videos, and 3D data), with the goal of developing and publicly providing effective attribute detection tools.

"Reshaping Datasets for Unsupervised Domain Adaptation", which is about our following works: [Gong et al., CVPR'12], [Gong, et al., ICML'13], and [Gong et al., NIPS'13].