Learning to Segment the Tail

04/02/2020
by   Xinting Hu, et al.
1

Real-world visual recognition requires handling the extreme sample imbalance in large-scale long-tailed data. We propose a "divide&conquer" strategy for the challenging LVIS task: divide the whole data into balanced parts and then apply incremental learning to conquer each one. This derives a novel learning paradigm: class-incremental few-shot learning, which is especially effective for the challenge evolving over time: 1) the class imbalance among the old-class knowledge review and 2) the few-shot data in new-class learning. We call our approach Learning to Segment the Tail (LST). In particular, we design an instance-level balanced replay scheme, which is a memory-efficient approximation to balance the instance-level samples from the old-class images. We also propose to use a meta-module for new-class learning, where the module parameters are shared across incremental phases, gaining the learning-to-learn knowledge incrementally, from the data-rich head to the data-poor tail. We empirically show that: at the expense of a little sacrifice of head-class forgetting, we can gain a significant 8.3% AP improvement for the tail classes with less than 10 instances, achieving an overall 2.0% AP boost for the whole 1,230 classes[%s].

READ FULL TEXT

page 4

page 8

research
06/07/2019

Learning Classifier Synthesis for Generalized Few-Shot Learning

Visual recognition in real-world requires handling long-tailed and even ...
research
08/03/2023

Neural Collapse Terminus: A Unified Solution for Class Incremental Learning and Its Variants

How to enable learnability for new classes while keeping the capability ...
research
03/31/2022

Few-Shot Class-Incremental Learning by Sampling Multi-Phase Tasks

New classes arise frequently in our ever-changing world, e.g., emerging ...
research
08/03/2023

Balanced Destruction-Reconstruction Dynamics for Memory-replay Class Incremental Learning

Class incremental learning (CIL) aims to incrementally update a trained ...
research
03/04/2019

Long-tail Relation Extraction via Knowledge Graph Embeddings and Graph Convolution Networks

We propose a distance supervised relation extraction approach for long-t...
research
07/27/2021

Co-Transport for Class-Incremental Learning

Traditional learning systems are trained in closed-world for a fixed num...
research
08/25/2023

Dual Compensation Residual Networks for Class Imbalanced Learning

Learning generalizable representation and classifier for class-imbalance...

Please sign up or login with your details

Forgot password? Click here to reset