Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287118244> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W4287118244 abstract "With the goal of making deep learning more label-efficient, a growing number of papers have been studying active learning (AL) for deep models. However, there are a number of issues in the prevalent experimental settings, mainly stemming from a lack of unified implementation and benchmarking. Issues in the current literature include sometimes contradictory observations on the performance of different AL algorithms, unintended exclusion of important generalization approaches such as data augmentation and SGD for optimization, a lack of study of evaluation facets like the labeling efficiency of AL, and little or no clarity on the scenarios in which AL outperforms random sampling (RS). In this work, we present a unified re-implementation of state-of-the-art AL algorithms in the context of image classification via our new open-source AL toolkit DISTIL, and we carefully study these issues as facets of effective evaluation. On the positive side, we show that AL techniques are $2times$ to $4times$ more label-efficient compared to RS with the use of data augmentation. Surprisingly, when data augmentation is included, there is no longer a consistent gain in using BADGE, a state-of-the-art approach, over simple uncertainty sampling. We then do a careful analysis of how existing approaches perform with varying amounts of redundancy and number of examples per class. Finally, we provide several insights for AL practitioners to consider in future work, such as the effect of the AL batch size, the effect of initialization, the importance of retraining the model at every round, and other insights." @default.
- W4287118244 created "2022-07-25" @default.
- W4287118244 creator A5000529247 @default.
- W4287118244 creator A5011703570 @default.
- W4287118244 creator A5042895651 @default.
- W4287118244 creator A5053292476 @default.
- W4287118244 creator A5089606464 @default.
- W4287118244 date "2021-06-16" @default.
- W4287118244 modified "2023-10-01" @default.
- W4287118244 title "Effective Evaluation of Deep Active Learning on Image Classification Tasks" @default.
- W4287118244 doi "https://doi.org/10.48550/arxiv.2106.15324" @default.
- W4287118244 hasPublicationYear "2021" @default.
- W4287118244 type Work @default.
- W4287118244 citedByCount "1" @default.
- W4287118244 countsByYear W42871182442023 @default.
- W4287118244 crossrefType "posted-content" @default.
- W4287118244 hasAuthorship W4287118244A5000529247 @default.
- W4287118244 hasAuthorship W4287118244A5011703570 @default.
- W4287118244 hasAuthorship W4287118244A5042895651 @default.
- W4287118244 hasAuthorship W4287118244A5053292476 @default.
- W4287118244 hasAuthorship W4287118244A5089606464 @default.
- W4287118244 hasBestOaLocation W42871182441 @default.
- W4287118244 hasConcept C108583219 @default.
- W4287118244 hasConcept C111919701 @default.
- W4287118244 hasConcept C114466953 @default.
- W4287118244 hasConcept C119857082 @default.
- W4287118244 hasConcept C134306372 @default.
- W4287118244 hasConcept C144133560 @default.
- W4287118244 hasConcept C151730666 @default.
- W4287118244 hasConcept C152124472 @default.
- W4287118244 hasConcept C154945302 @default.
- W4287118244 hasConcept C155202549 @default.
- W4287118244 hasConcept C162853370 @default.
- W4287118244 hasConcept C177148314 @default.
- W4287118244 hasConcept C185592680 @default.
- W4287118244 hasConcept C199360897 @default.
- W4287118244 hasConcept C2777146004 @default.
- W4287118244 hasConcept C2778712577 @default.
- W4287118244 hasConcept C2779343474 @default.
- W4287118244 hasConcept C33923547 @default.
- W4287118244 hasConcept C41008148 @default.
- W4287118244 hasConcept C55493867 @default.
- W4287118244 hasConcept C86251818 @default.
- W4287118244 hasConcept C86803240 @default.
- W4287118244 hasConceptScore W4287118244C108583219 @default.
- W4287118244 hasConceptScore W4287118244C111919701 @default.
- W4287118244 hasConceptScore W4287118244C114466953 @default.
- W4287118244 hasConceptScore W4287118244C119857082 @default.
- W4287118244 hasConceptScore W4287118244C134306372 @default.
- W4287118244 hasConceptScore W4287118244C144133560 @default.
- W4287118244 hasConceptScore W4287118244C151730666 @default.
- W4287118244 hasConceptScore W4287118244C152124472 @default.
- W4287118244 hasConceptScore W4287118244C154945302 @default.
- W4287118244 hasConceptScore W4287118244C155202549 @default.
- W4287118244 hasConceptScore W4287118244C162853370 @default.
- W4287118244 hasConceptScore W4287118244C177148314 @default.
- W4287118244 hasConceptScore W4287118244C185592680 @default.
- W4287118244 hasConceptScore W4287118244C199360897 @default.
- W4287118244 hasConceptScore W4287118244C2777146004 @default.
- W4287118244 hasConceptScore W4287118244C2778712577 @default.
- W4287118244 hasConceptScore W4287118244C2779343474 @default.
- W4287118244 hasConceptScore W4287118244C33923547 @default.
- W4287118244 hasConceptScore W4287118244C41008148 @default.
- W4287118244 hasConceptScore W4287118244C55493867 @default.
- W4287118244 hasConceptScore W4287118244C86251818 @default.
- W4287118244 hasConceptScore W4287118244C86803240 @default.
- W4287118244 hasLocation W42871182441 @default.
- W4287118244 hasOpenAccess W4287118244 @default.
- W4287118244 hasPrimaryLocation W42871182441 @default.
- W4287118244 hasRelatedWork W2765476116 @default.
- W4287118244 hasRelatedWork W2806031239 @default.
- W4287118244 hasRelatedWork W3014300295 @default.
- W4287118244 hasRelatedWork W4223943233 @default.
- W4287118244 hasRelatedWork W4225161397 @default.
- W4287118244 hasRelatedWork W4309045103 @default.
- W4287118244 hasRelatedWork W4312200629 @default.
- W4287118244 hasRelatedWork W4360585206 @default.
- W4287118244 hasRelatedWork W4364306694 @default.
- W4287118244 hasRelatedWork W4380086463 @default.
- W4287118244 isParatext "false" @default.
- W4287118244 isRetracted "false" @default.
- W4287118244 workType "article" @default.