Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287117310> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W4287117310 abstract "In many practical applications of machine learning data arrives sequentially over time in large chunks. Practitioners have then to decide how to allocate their computational budget in order to obtain the best performance at any point in time. Online learning theory for convex optimization suggests that the best strategy is to use data as soon as it arrives. However, this might not be the best strategy when using deep non-linear networks, particularly when these perform multiple passes over each chunk of data rendering the overall distribution non i.i.d.. In this paper, we formalize this learning setting in the simplest scenario in which each data chunk is drawn from the same underlying distribution, and make a first attempt at empirically answering the following questions: How long should the learner wait before training on the newly arrived chunks? What architecture should the learner adopt? Should the learner increase capacity over time as more data is observed? We probe this learning setting using convolutional neural networks trained on classic computer vision benchmarks as well as a large transformer model trained on a large-scale language modeling task. Code is available at url{www.github.com/facebookresearch/ALMA}." @default.
- W4287117310 created "2022-07-25" @default.
- W4287117310 creator A5027101473 @default.
- W4287117310 creator A5035384485 @default.
- W4287117310 creator A5038906848 @default.
- W4287117310 creator A5076248976 @default.
- W4287117310 creator A5077887756 @default.
- W4287117310 date "2021-06-17" @default.
- W4287117310 modified "2023-09-24" @default.
- W4287117310 title "On Anytime Learning at Macroscale" @default.
- W4287117310 doi "https://doi.org/10.48550/arxiv.2106.09563" @default.
- W4287117310 hasPublicationYear "2021" @default.
- W4287117310 type Work @default.
- W4287117310 citedByCount "0" @default.
- W4287117310 crossrefType "posted-content" @default.
- W4287117310 hasAuthorship W4287117310A5027101473 @default.
- W4287117310 hasAuthorship W4287117310A5035384485 @default.
- W4287117310 hasAuthorship W4287117310A5038906848 @default.
- W4287117310 hasAuthorship W4287117310A5076248976 @default.
- W4287117310 hasAuthorship W4287117310A5077887756 @default.
- W4287117310 hasBestOaLocation W42871173101 @default.
- W4287117310 hasConcept C108583219 @default.
- W4287117310 hasConcept C119857082 @default.
- W4287117310 hasConcept C121332964 @default.
- W4287117310 hasConcept C123657996 @default.
- W4287117310 hasConcept C137293760 @default.
- W4287117310 hasConcept C142362112 @default.
- W4287117310 hasConcept C153349607 @default.
- W4287117310 hasConcept C154945302 @default.
- W4287117310 hasConcept C162324750 @default.
- W4287117310 hasConcept C165801399 @default.
- W4287117310 hasConcept C187736073 @default.
- W4287117310 hasConcept C205711294 @default.
- W4287117310 hasConcept C2780451532 @default.
- W4287117310 hasConcept C28006648 @default.
- W4287117310 hasConcept C41008148 @default.
- W4287117310 hasConcept C62520636 @default.
- W4287117310 hasConcept C66322947 @default.
- W4287117310 hasConcept C81363708 @default.
- W4287117310 hasConceptScore W4287117310C108583219 @default.
- W4287117310 hasConceptScore W4287117310C119857082 @default.
- W4287117310 hasConceptScore W4287117310C121332964 @default.
- W4287117310 hasConceptScore W4287117310C123657996 @default.
- W4287117310 hasConceptScore W4287117310C137293760 @default.
- W4287117310 hasConceptScore W4287117310C142362112 @default.
- W4287117310 hasConceptScore W4287117310C153349607 @default.
- W4287117310 hasConceptScore W4287117310C154945302 @default.
- W4287117310 hasConceptScore W4287117310C162324750 @default.
- W4287117310 hasConceptScore W4287117310C165801399 @default.
- W4287117310 hasConceptScore W4287117310C187736073 @default.
- W4287117310 hasConceptScore W4287117310C205711294 @default.
- W4287117310 hasConceptScore W4287117310C2780451532 @default.
- W4287117310 hasConceptScore W4287117310C28006648 @default.
- W4287117310 hasConceptScore W4287117310C41008148 @default.
- W4287117310 hasConceptScore W4287117310C62520636 @default.
- W4287117310 hasConceptScore W4287117310C66322947 @default.
- W4287117310 hasConceptScore W4287117310C81363708 @default.
- W4287117310 hasLocation W42871173101 @default.
- W4287117310 hasOpenAccess W4287117310 @default.
- W4287117310 hasPrimaryLocation W42871173101 @default.
- W4287117310 hasRelatedWork W1499860779 @default.
- W4287117310 hasRelatedWork W2896257747 @default.
- W4287117310 hasRelatedWork W2951576432 @default.
- W4287117310 hasRelatedWork W2961085424 @default.
- W4287117310 hasRelatedWork W2963218179 @default.
- W4287117310 hasRelatedWork W3129403840 @default.
- W4287117310 hasRelatedWork W3211971560 @default.
- W4287117310 hasRelatedWork W335022172 @default.
- W4287117310 hasRelatedWork W41957589 @default.
- W4287117310 hasRelatedWork W4287811188 @default.
- W4287117310 isParatext "false" @default.
- W4287117310 isRetracted "false" @default.
- W4287117310 workType "article" @default.