SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.14042 · CVPR · 2023

Class-Incremental Exemplar Compression for Class-Incremental Learning

Bernt Schiele, Qianru Sun, Yaoyao Liu, Zilin Luo

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Exemplar-based class-incremental learning (CIL) [36] finetunes the model with all samples of new classes but fewshot exemplars of old classes in each incremental phase, where the "few-shot" abides by the limited memory budget. In this paper, we break this "few-shot" limit based on a simple yet surprisingly effective idea: compressing exemplars by downsampling non-discriminative pixels and saving "many-shot" compressed exemplars in the memory. Without needing any manual annotation, we achieve this compression by generating 0-1 masks on discriminative pixels from class activation maps (CAM) [49]. We propose an adaptive mask generation model called classincremental masking (CIM) to explicitly resolve two difficulties of using CAM: 1) transforming the heatmaps of CAM to 0-1 masks with an arbitrary threshold leads to a trade-off between the coverage on discriminative pixels and the quantity of exemplars, as the total memory is fixed; and 2) optimal thresholds vary for different object classes, which is particularly obvious in the dynamic environment of CIL. We optimize the CIM model alternatively with the conventional CIL model through a bilevel optimization problem [40]. We conduct extensive experiments on high-resolution CIL benchmarks including Food-101, ImageNet-100, and ImageNet-1000, and show that using the compressed exemplars by CIM can achieve a new stateof-the-art CIL accuracy, e.g., 4.8 percentage points higher than FOSTER [42] on 10-Phase ImageNet-1000. Our code is available at https://github.com/xfflzl/CIM-CIL.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2303.14042")
get_code_for_paper("2303.14042")
have("2303.14042")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2303.14042.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2303.14042.svg)](https://syntology.ai/paper/2303.14042)

Connect an agent — have() is free.