Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387171859> ?p ?o ?g. }
Showing items 1 to 65 of
65
with 100 items per page.
- W4387171859 abstract "Models that can actively seek out the best quality training data hold the promise of more accurate, adaptable, and efficient machine learning. Active learning techniques often tend to prefer examples that are the most difficult to classify. While this works well on homogeneous datasets, we find that it can lead to catastrophic failures when performed on multiple distributions with different degrees of label noise or heteroskedasticity. These active learning algorithms strongly prefer to draw from the distribution with more noise, even if their examples have no informative structure (such as solid color images with random labels). To this end, we demonstrate the catastrophic failure of these active learning algorithms on heteroskedastic distributions and propose a fine-tuning-based approach to mitigate these failures. Further, we propose a new algorithm that incorporates a model difference scoring function for each data point to filter out the noisy examples and sample clean examples that maximize accuracy, outperforming the existing active learning techniques on the heteroskedastic datasets. We hope these observations and techniques are immediately helpful to practitioners and can help to challenge common assumptions in the design of active learning algorithms. Our code is available at https://github.com/savya08/Active-Learning-on-Heteroskedastic-Distributions." @default.
- W4387171859 created "2023-09-30" @default.
- W4387171859 creator A5003184366 @default.
- W4387171859 creator A5030891292 @default.
- W4387171859 creator A5043228083 @default.
- W4387171859 creator A5046423882 @default.
- W4387171859 creator A5063681225 @default.
- W4387171859 creator A5066563664 @default.
- W4387171859 date "2023-09-28" @default.
- W4387171859 modified "2023-09-30" @default.
- W4387171859 title "Understanding and Improving Neural Active Learning on Heteroskedastic Distributions" @default.
- W4387171859 doi "https://doi.org/10.3233/faia230402" @default.
- W4387171859 hasPublicationYear "2023" @default.
- W4387171859 type Work @default.
- W4387171859 citedByCount "0" @default.
- W4387171859 crossrefType "book-chapter" @default.
- W4387171859 hasAuthorship W4387171859A5003184366 @default.
- W4387171859 hasAuthorship W4387171859A5030891292 @default.
- W4387171859 hasAuthorship W4387171859A5043228083 @default.
- W4387171859 hasAuthorship W4387171859A5046423882 @default.
- W4387171859 hasAuthorship W4387171859A5063681225 @default.
- W4387171859 hasAuthorship W4387171859A5066563664 @default.
- W4387171859 hasBestOaLocation W43871718591 @default.
- W4387171859 hasConcept C101104100 @default.
- W4387171859 hasConcept C106131492 @default.
- W4387171859 hasConcept C115961682 @default.
- W4387171859 hasConcept C119857082 @default.
- W4387171859 hasConcept C124101348 @default.
- W4387171859 hasConcept C154945302 @default.
- W4387171859 hasConcept C2524010 @default.
- W4387171859 hasConcept C28719098 @default.
- W4387171859 hasConcept C31972630 @default.
- W4387171859 hasConcept C33923547 @default.
- W4387171859 hasConcept C41008148 @default.
- W4387171859 hasConcept C77967617 @default.
- W4387171859 hasConcept C99498987 @default.
- W4387171859 hasConceptScore W4387171859C101104100 @default.
- W4387171859 hasConceptScore W4387171859C106131492 @default.
- W4387171859 hasConceptScore W4387171859C115961682 @default.
- W4387171859 hasConceptScore W4387171859C119857082 @default.
- W4387171859 hasConceptScore W4387171859C124101348 @default.
- W4387171859 hasConceptScore W4387171859C154945302 @default.
- W4387171859 hasConceptScore W4387171859C2524010 @default.
- W4387171859 hasConceptScore W4387171859C28719098 @default.
- W4387171859 hasConceptScore W4387171859C31972630 @default.
- W4387171859 hasConceptScore W4387171859C33923547 @default.
- W4387171859 hasConceptScore W4387171859C41008148 @default.
- W4387171859 hasConceptScore W4387171859C77967617 @default.
- W4387171859 hasConceptScore W4387171859C99498987 @default.
- W4387171859 hasLocation W43871718591 @default.
- W4387171859 hasOpenAccess W4387171859 @default.
- W4387171859 hasPrimaryLocation W43871718591 @default.
- W4387171859 hasRelatedWork W2597787948 @default.
- W4387171859 hasRelatedWork W2954428433 @default.
- W4387171859 hasRelatedWork W3025582806 @default.
- W4387171859 hasRelatedWork W3047894882 @default.
- W4387171859 hasRelatedWork W3136151706 @default.
- W4387171859 hasRelatedWork W3177723589 @default.
- W4387171859 hasRelatedWork W3196155444 @default.
- W4387171859 hasRelatedWork W3208584567 @default.
- W4387171859 hasRelatedWork W4320063314 @default.
- W4387171859 hasRelatedWork W4366320140 @default.
- W4387171859 isParatext "false" @default.
- W4387171859 isRetracted "false" @default.
- W4387171859 workType "book-chapter" @default.