DiAPR: Dimensionally-Allocated Prototype Refinement for Non-Exemplar Class Incremental Learning.

Ruixuan Gao, Qijun Zhao, Keren Fu

Browse the full AAAI paper archive.