Matches in SemOpenAlex for { <https://semopenalex.org/work/W4322717029> ?p ?o ?g. }
Showing items 1 to 87 of
87
with 100 items per page.
- W4322717029 abstract "This paper proposes a novel, efficient transfer learning method, called Scalable Weight Reparametrization (SWR) that is efficient and effective for multiple downstream tasks. Efficient transfer learning involves utilizing a pre-trained model trained on a larger dataset and repurposing it for downstream tasks with the aim of maximizing the reuse of the pre-trained model. However, previous works have led to an increase in updated parameters and task-specific modules, resulting in more computations, especially for tiny models. Additionally, there has been no practical consideration for controlling the number of updated parameters. To address these issues, we suggest learning a policy network that can decide where to reparametrize the pre-trained model, while adhering to a given constraint for the number of updated parameters. The policy network is only used during the transfer learning process and not afterward. As a result, our approach attains state-of-the-art performance in a proposed multi-lingual keyword spotting and a standard benchmark, ImageNet-to-Sketch, while requiring zero additional computations and significantly fewer additional parameters." @default.
- W4322717029 created "2023-03-03" @default.
- W4322717029 creator A5018347257 @default.
- W4322717029 creator A5027266141 @default.
- W4322717029 creator A5039155214 @default.
- W4322717029 creator A5057566014 @default.
- W4322717029 date "2023-02-26" @default.
- W4322717029 modified "2023-09-29" @default.
- W4322717029 title "Scalable Weight Reparametrization for Efficient Transfer Learning" @default.
- W4322717029 doi "https://doi.org/10.48550/arxiv.2302.13435" @default.
- W4322717029 hasPublicationYear "2023" @default.
- W4322717029 type Work @default.
- W4322717029 citedByCount "0" @default.
- W4322717029 crossrefType "posted-content" @default.
- W4322717029 hasAuthorship W4322717029A5018347257 @default.
- W4322717029 hasAuthorship W4322717029A5027266141 @default.
- W4322717029 hasAuthorship W4322717029A5039155214 @default.
- W4322717029 hasAuthorship W4322717029A5057566014 @default.
- W4322717029 hasBestOaLocation W43227170291 @default.
- W4322717029 hasConcept C111919701 @default.
- W4322717029 hasConcept C11413529 @default.
- W4322717029 hasConcept C119857082 @default.
- W4322717029 hasConcept C127413603 @default.
- W4322717029 hasConcept C13280743 @default.
- W4322717029 hasConcept C150899416 @default.
- W4322717029 hasConcept C154945302 @default.
- W4322717029 hasConcept C162324750 @default.
- W4322717029 hasConcept C185798385 @default.
- W4322717029 hasConcept C187736073 @default.
- W4322717029 hasConcept C18903297 @default.
- W4322717029 hasConcept C205649164 @default.
- W4322717029 hasConcept C206588197 @default.
- W4322717029 hasConcept C21547014 @default.
- W4322717029 hasConcept C2776036281 @default.
- W4322717029 hasConcept C2776207758 @default.
- W4322717029 hasConcept C2779231336 @default.
- W4322717029 hasConcept C2780451532 @default.
- W4322717029 hasConcept C41008148 @default.
- W4322717029 hasConcept C45374587 @default.
- W4322717029 hasConcept C48044578 @default.
- W4322717029 hasConcept C519536355 @default.
- W4322717029 hasConcept C77088390 @default.
- W4322717029 hasConcept C78519656 @default.
- W4322717029 hasConcept C86803240 @default.
- W4322717029 hasConcept C98045186 @default.
- W4322717029 hasConceptScore W4322717029C111919701 @default.
- W4322717029 hasConceptScore W4322717029C11413529 @default.
- W4322717029 hasConceptScore W4322717029C119857082 @default.
- W4322717029 hasConceptScore W4322717029C127413603 @default.
- W4322717029 hasConceptScore W4322717029C13280743 @default.
- W4322717029 hasConceptScore W4322717029C150899416 @default.
- W4322717029 hasConceptScore W4322717029C154945302 @default.
- W4322717029 hasConceptScore W4322717029C162324750 @default.
- W4322717029 hasConceptScore W4322717029C185798385 @default.
- W4322717029 hasConceptScore W4322717029C187736073 @default.
- W4322717029 hasConceptScore W4322717029C18903297 @default.
- W4322717029 hasConceptScore W4322717029C205649164 @default.
- W4322717029 hasConceptScore W4322717029C206588197 @default.
- W4322717029 hasConceptScore W4322717029C21547014 @default.
- W4322717029 hasConceptScore W4322717029C2776036281 @default.
- W4322717029 hasConceptScore W4322717029C2776207758 @default.
- W4322717029 hasConceptScore W4322717029C2779231336 @default.
- W4322717029 hasConceptScore W4322717029C2780451532 @default.
- W4322717029 hasConceptScore W4322717029C41008148 @default.
- W4322717029 hasConceptScore W4322717029C45374587 @default.
- W4322717029 hasConceptScore W4322717029C48044578 @default.
- W4322717029 hasConceptScore W4322717029C519536355 @default.
- W4322717029 hasConceptScore W4322717029C77088390 @default.
- W4322717029 hasConceptScore W4322717029C78519656 @default.
- W4322717029 hasConceptScore W4322717029C86803240 @default.
- W4322717029 hasConceptScore W4322717029C98045186 @default.
- W4322717029 hasLocation W43227170291 @default.
- W4322717029 hasOpenAccess W4322717029 @default.
- W4322717029 hasPrimaryLocation W43227170291 @default.
- W4322717029 hasRelatedWork W112744582 @default.
- W4322717029 hasRelatedWork W1992807924 @default.
- W4322717029 hasRelatedWork W2152087638 @default.
- W4322717029 hasRelatedWork W2375463041 @default.
- W4322717029 hasRelatedWork W2521990067 @default.
- W4322717029 hasRelatedWork W2789601449 @default.
- W4322717029 hasRelatedWork W3018421652 @default.
- W4322717029 hasRelatedWork W3088526462 @default.
- W4322717029 hasRelatedWork W4281382123 @default.
- W4322717029 hasRelatedWork W4308262314 @default.
- W4322717029 isParatext "false" @default.
- W4322717029 isRetracted "false" @default.
- W4322717029 workType "article" @default.