Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387323232> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W4387323232 abstract "Continual learning is inherently a constrained learning problem. The goal is to learn a predictor under a emph{no-forgetting} requirement. Although several prior studies formulate it as such, they do not solve the constrained problem explicitly. In this work, we show that it is both possible and beneficial to undertake the constrained optimization problem directly. To do this, we leverage recent results in constrained learning through Lagrangian duality. We focus on memory-based methods, where a small subset of samples from previous tasks can be stored in a replay buffer. In this setting, we analyze two versions of the continual learning problem: a coarse approach with constraints at the task level and a fine approach with constraints at the sample level. We show that dual variables indicate the sensitivity of the optimal value with respect to constraint perturbations. We then leverage this result to partition the buffer in the coarse approach, allocating more resources to harder tasks, and to populate the buffer in the fine approach, including only impactful samples. We derive sub-optimality bounds, and empirically corroborate our theoretical results in various continual learning benchmarks. We also discuss the limitations of these methods with respect to the amount of memory available and the number of constraints involved in the optimization problem." @default.
- W4387323232 created "2023-10-04" @default.
- W4387323232 creator A5024099070 @default.
- W4387323232 creator A5059310658 @default.
- W4387323232 creator A5065389164 @default.
- W4387323232 creator A5078862959 @default.
- W4387323232 date "2023-09-29" @default.
- W4387323232 modified "2023-10-05" @default.
- W4387323232 title "Primal-Dual Continual Learning: Stability and Plasticity through Lagrange Multipliers" @default.
- W4387323232 doi "https://doi.org/10.48550/arxiv.2310.00154" @default.
- W4387323232 hasPublicationYear "2023" @default.
- W4387323232 type Work @default.
- W4387323232 citedByCount "0" @default.
- W4387323232 crossrefType "posted-content" @default.
- W4387323232 hasAuthorship W4387323232A5024099070 @default.
- W4387323232 hasAuthorship W4387323232A5059310658 @default.
- W4387323232 hasAuthorship W4387323232A5065389164 @default.
- W4387323232 hasAuthorship W4387323232A5078862959 @default.
- W4387323232 hasBestOaLocation W43873232321 @default.
- W4387323232 hasConcept C112972136 @default.
- W4387323232 hasConcept C119857082 @default.
- W4387323232 hasConcept C124952713 @default.
- W4387323232 hasConcept C126255220 @default.
- W4387323232 hasConcept C137836250 @default.
- W4387323232 hasConcept C138885662 @default.
- W4387323232 hasConcept C142362112 @default.
- W4387323232 hasConcept C153083717 @default.
- W4387323232 hasConcept C154945302 @default.
- W4387323232 hasConcept C2524010 @default.
- W4387323232 hasConcept C2776036281 @default.
- W4387323232 hasConcept C2780980858 @default.
- W4387323232 hasConcept C2989514635 @default.
- W4387323232 hasConcept C33923547 @default.
- W4387323232 hasConcept C41008148 @default.
- W4387323232 hasConcept C41895202 @default.
- W4387323232 hasConcept C7149132 @default.
- W4387323232 hasConcept C73684929 @default.
- W4387323232 hasConceptScore W4387323232C112972136 @default.
- W4387323232 hasConceptScore W4387323232C119857082 @default.
- W4387323232 hasConceptScore W4387323232C124952713 @default.
- W4387323232 hasConceptScore W4387323232C126255220 @default.
- W4387323232 hasConceptScore W4387323232C137836250 @default.
- W4387323232 hasConceptScore W4387323232C138885662 @default.
- W4387323232 hasConceptScore W4387323232C142362112 @default.
- W4387323232 hasConceptScore W4387323232C153083717 @default.
- W4387323232 hasConceptScore W4387323232C154945302 @default.
- W4387323232 hasConceptScore W4387323232C2524010 @default.
- W4387323232 hasConceptScore W4387323232C2776036281 @default.
- W4387323232 hasConceptScore W4387323232C2780980858 @default.
- W4387323232 hasConceptScore W4387323232C2989514635 @default.
- W4387323232 hasConceptScore W4387323232C33923547 @default.
- W4387323232 hasConceptScore W4387323232C41008148 @default.
- W4387323232 hasConceptScore W4387323232C41895202 @default.
- W4387323232 hasConceptScore W4387323232C7149132 @default.
- W4387323232 hasConceptScore W4387323232C73684929 @default.
- W4387323232 hasLocation W43873232321 @default.
- W4387323232 hasOpenAccess W4387323232 @default.
- W4387323232 hasPrimaryLocation W43873232321 @default.
- W4387323232 hasRelatedWork W2001953799 @default.
- W4387323232 hasRelatedWork W2015060724 @default.
- W4387323232 hasRelatedWork W2093559692 @default.
- W4387323232 hasRelatedWork W2100183008 @default.
- W4387323232 hasRelatedWork W2350481022 @default.
- W4387323232 hasRelatedWork W2355999325 @default.
- W4387323232 hasRelatedWork W2409674212 @default.
- W4387323232 hasRelatedWork W2806604749 @default.
- W4387323232 hasRelatedWork W318488473 @default.
- W4387323232 hasRelatedWork W3209065746 @default.
- W4387323232 isParatext "false" @default.
- W4387323232 isRetracted "false" @default.
- W4387323232 workType "article" @default.