Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385767824> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W4385767824 abstract "Despite deep learning has achieved great success, it often relies on a large amount of training data with accurate labels, which are expensive and time-consuming to collect. A prominent direction to reduce the cost is to learn with noisy labels, which are ubiquitous in the real-world applications. A critical challenge for such a learning task is to reduce the effect of network memorization on the falsely-labeled data. In this work, we propose an iterative selection approach based on the Weibull mixture model, which identifies clean data by considering the overall learning dynamics of each data instance. In contrast to the previous small-loss heuristics, we leverage the observation that deep network is easy to memorize and hard to forget clean data. In particular, we measure the difficulty of memorization and forgetting for each instance via the transition times between being misclassified and being memorized in training, and integrate them into a novel metric for selection. Based on the proposed metric, we retain a subset of identified clean data and repeat the selection procedure to iteratively refine the clean subset, which is finally used for model training. To validate our method, we perform extensive experiments on synthetic noisy datasets and real-world web data, and our strategy outperforms existing noisy-label learning methods." @default.
- W4385767824 created "2023-08-12" @default.
- W4385767824 creator A5015970030 @default.
- W4385767824 creator A5062881932 @default.
- W4385767824 creator A5074627818 @default.
- W4385767824 creator A5088442800 @default.
- W4385767824 date "2023-08-01" @default.
- W4385767824 modified "2023-09-28" @default.
- W4385767824 title "MILD: Modeling the Instance Learning Dynamics for Learning with Noisy Labels" @default.
- W4385767824 doi "https://doi.org/10.24963/ijcai.2023/92" @default.
- W4385767824 hasPublicationYear "2023" @default.
- W4385767824 type Work @default.
- W4385767824 citedByCount "0" @default.
- W4385767824 crossrefType "proceedings-article" @default.
- W4385767824 hasAuthorship W4385767824A5015970030 @default.
- W4385767824 hasAuthorship W4385767824A5062881932 @default.
- W4385767824 hasAuthorship W4385767824A5074627818 @default.
- W4385767824 hasAuthorship W4385767824A5088442800 @default.
- W4385767824 hasBestOaLocation W43857678241 @default.
- W4385767824 hasConcept C108583219 @default.
- W4385767824 hasConcept C111919701 @default.
- W4385767824 hasConcept C119857082 @default.
- W4385767824 hasConcept C124101348 @default.
- W4385767824 hasConcept C127705205 @default.
- W4385767824 hasConcept C138885662 @default.
- W4385767824 hasConcept C145420912 @default.
- W4385767824 hasConcept C153083717 @default.
- W4385767824 hasConcept C154945302 @default.
- W4385767824 hasConcept C162324750 @default.
- W4385767824 hasConcept C176217482 @default.
- W4385767824 hasConcept C187736073 @default.
- W4385767824 hasConcept C21547014 @default.
- W4385767824 hasConcept C2780451532 @default.
- W4385767824 hasConcept C30038468 @default.
- W4385767824 hasConcept C33923547 @default.
- W4385767824 hasConcept C41008148 @default.
- W4385767824 hasConcept C41895202 @default.
- W4385767824 hasConcept C7149132 @default.
- W4385767824 hasConcept C93959086 @default.
- W4385767824 hasConcept C95623464 @default.
- W4385767824 hasConceptScore W4385767824C108583219 @default.
- W4385767824 hasConceptScore W4385767824C111919701 @default.
- W4385767824 hasConceptScore W4385767824C119857082 @default.
- W4385767824 hasConceptScore W4385767824C124101348 @default.
- W4385767824 hasConceptScore W4385767824C127705205 @default.
- W4385767824 hasConceptScore W4385767824C138885662 @default.
- W4385767824 hasConceptScore W4385767824C145420912 @default.
- W4385767824 hasConceptScore W4385767824C153083717 @default.
- W4385767824 hasConceptScore W4385767824C154945302 @default.
- W4385767824 hasConceptScore W4385767824C162324750 @default.
- W4385767824 hasConceptScore W4385767824C176217482 @default.
- W4385767824 hasConceptScore W4385767824C187736073 @default.
- W4385767824 hasConceptScore W4385767824C21547014 @default.
- W4385767824 hasConceptScore W4385767824C2780451532 @default.
- W4385767824 hasConceptScore W4385767824C30038468 @default.
- W4385767824 hasConceptScore W4385767824C33923547 @default.
- W4385767824 hasConceptScore W4385767824C41008148 @default.
- W4385767824 hasConceptScore W4385767824C41895202 @default.
- W4385767824 hasConceptScore W4385767824C7149132 @default.
- W4385767824 hasConceptScore W4385767824C93959086 @default.
- W4385767824 hasConceptScore W4385767824C95623464 @default.
- W4385767824 hasLocation W43857678241 @default.
- W4385767824 hasLocation W43857678242 @default.
- W4385767824 hasOpenAccess W4385767824 @default.
- W4385767824 hasPrimaryLocation W43857678241 @default.
- W4385767824 hasRelatedWork W2795261237 @default.
- W4385767824 hasRelatedWork W3014300295 @default.
- W4385767824 hasRelatedWork W3164822677 @default.
- W4385767824 hasRelatedWork W4223943233 @default.
- W4385767824 hasRelatedWork W4225161397 @default.
- W4385767824 hasRelatedWork W4312200629 @default.
- W4385767824 hasRelatedWork W4360585206 @default.
- W4385767824 hasRelatedWork W4364306694 @default.
- W4385767824 hasRelatedWork W4380075502 @default.
- W4385767824 hasRelatedWork W4380086463 @default.
- W4385767824 isParatext "false" @default.
- W4385767824 isRetracted "false" @default.
- W4385767824 workType "article" @default.