• Publications
  • Influence
A Method of Converting Z-number to Classical Fuzzy Number
The notion Z-number introduced by Zadeh in 2011 has more capability to describe the uncertain information. Now that the theories about Z-number is not mature, how to convert Z-number to classicalExpand
  • 149
  • 25
Few-Shot Object Detection via Feature Reweighting
In this work we develop a few-shot object detector that can learn to detect novel objects from only a few annotated examples. Expand
  • 81
  • 22
  • PDF
Decoupling Representation and Classifier for Long-Tailed Recognition
The long-tail distribution of the visual world poses great challenges for deep learning based classification models on how to handle the class imbalance problem. Expand
  • 50
  • 16
  • PDF
Decision Making Using Z-numbers under Uncertain Environment
Multi-criteria decision making (MCDM) under uncertain environment is still an open issue. Recently, Znumber has been developed by Zadeh to model fuzzy numbers with the confidence degree. In thisExpand
  • 101
  • 14
Policy Optimization with Demonstrations
We propose to effectively leverage available demonstrations to guide exploration through enforcing occupancy measure matching between the learned policy and current demonstrations, and develop a novel Policy Optimization from Demonstration (POfD) method. Expand
  • 45
  • 9
  • PDF
Conflict management based on belief function entropy in sensor fusion
A weighted averaging combination method on basis of evidence distance and Deng entropy is brought up to manage conflict in sensor data fusion. Expand
  • 34
  • 4
Stable strategies analysis based on the utility of Z-number in the evolutionary games
Z-number is combined with “restriction” and “reliability”, which is an efficient framework to simulate the thinking of human. Expand
  • 91
  • 3
An Improved Deng Entropy and Its Application in Pattern Recognition
In this paper, a modified function is proposed by considering the scale of the frame of discernment and the influence of the intersection between statements on uncertainty. Expand
  • 60
  • 3
Overcoming Classifier Imbalance for Long-Tail Object Detection With Balanced Group Softmax
  • Y. Li, Tao Wang, +4 authors Jiashi Feng
  • Computer Science, Mathematics
  • IEEE/CVF Conference on Computer Vision and…
  • 1 June 2020
We propose a novel balanced group softmax (BAGS) module for balancing the classifiers within the detection frameworks through group-wise training. Expand
  • 12
  • 3
  • PDF
The Devil is in Classification: A Simple Framework for Long-tail Instance Segmentation
We systematically investigate performance drop of the state-of-the-art two-stage instance segmentation model Mask R-CNN on the recent long-tail LVIS dataset, and unveil that a major cause is the inaccurate classification of object proposals. Expand
  • 5
  • 3