Article Source

• Heterogeneous Continual Learning

Heterogeneous Continual Learning

• CVPR 2023 Highlight: Heterogeneous Continual Learning by Divyam Madaan (NYU, NVIDIA), Hongxu Yin (NVIDIA), Wonmin Byeon (NVIDIA), Jan Kautz (NVIDIA), and Pavlo Molchanov (NVIDIA).

Abstract

We propose a novel framework and a solution to tackle the continual learning (CL) problem with changing network architectures. Additionally, we consider a setup of limited access to previous data and propose Quick Deep Inversion (QDI) to recover prior task visual features to support knowledge transfer.