Timezone: »

Optimal Continual Learning has Perfect Memory and is NP-hard
Jeremias Knoblauch · Hisham Husain · Tom Diethe

Wed Jul 15 12:00 PM -- 12:45 PM & Thu Jul 16 01:00 AM -- 01:45 AM (PDT) @

Continual Learning (CL) algorithms incrementally learn a predictor or representation across multiple sequentially observed tasks. Designing CL algorithms that perform reliably and avoid so-called catastrophic forgetting has proven a persistent challenge. The current paper develops a theoretical approach that explains why. In particular, we derive the computational properties which CL algorithms would have to possess in order to avoid catastrophic forgetting. Our main finding is that such optimal CL algorithms generally solve an NP-hard problem and will require perfect memory to do so. The findings are of theoretical interest, but also explain the excellent performance of CL algorithms using experience replay, episodic memory and core sets relative to regularization-based approaches.

Author Information

Jeremias Knoblauch (Warwick University & The Alan Turing Institute)
Hisham Husain (Australian National University)
Tom Diethe (Amazon)

More from the Same Authors