Matches in SemOpenAlex for { <https://semopenalex.org/work/W3034454003> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W3034454003 endingPage "3291" @default.
- W3034454003 startingPage "3280" @default.
- W3034454003 abstract "Weak supervision is a popular method for building machine learning models without relying on ground truth annotations. Instead, it generates probabilistic training labels by estimating the accuracies of multiple noisy labeling sources (e.g., heuristics, crowd workers). Existing approaches use latent variable estimation to model the noisy sources, but these methods can be computationally expensive, scaling superlinearly in the data. In this work, we show that, for a class of latent variable models highly applicable to weak supervision, we can find a closed-form solution to model parameters, obviating the need for iterative solutions like stochastic gradient descent (SGD). We use this insight to build FlyingSquid, a weak supervision framework that runs orders of magnitude faster than previous weak supervision approaches and requires fewer assumptions. In particular, we prove bounds on generalization error without assuming that the latent variable model can exactly parameterize the underlying data distribution. Empirically, we validate FlyingSquid on benchmark weak supervision datasets and find that it achieves the same or higher quality compared to previous approaches without the need to tune an SGD procedure, recovers model parameters 170 times faster on average, and enables new video analysis and online learning applications." @default.
- W3034454003 created "2020-06-19" @default.
- W3034454003 creator A5026912066 @default.
- W3034454003 creator A5032865467 @default.
- W3034454003 creator A5037444018 @default.
- W3034454003 creator A5039665431 @default.
- W3034454003 creator A5060745338 @default.
- W3034454003 creator A5070267604 @default.
- W3034454003 date "2020-07-12" @default.
- W3034454003 modified "2023-09-23" @default.
- W3034454003 title "Fast and Three-rious: Speeding Up Weak Supervision with Triplet Methods" @default.
- W3034454003 hasPublicationYear "2020" @default.
- W3034454003 type Work @default.
- W3034454003 sameAs 3034454003 @default.
- W3034454003 citedByCount "15" @default.
- W3034454003 countsByYear W30344540032020 @default.
- W3034454003 countsByYear W30344540032021 @default.
- W3034454003 crossrefType "proceedings-article" @default.
- W3034454003 hasAuthorship W3034454003A5026912066 @default.
- W3034454003 hasAuthorship W3034454003A5032865467 @default.
- W3034454003 hasAuthorship W3034454003A5037444018 @default.
- W3034454003 hasAuthorship W3034454003A5039665431 @default.
- W3034454003 hasAuthorship W3034454003A5060745338 @default.
- W3034454003 hasAuthorship W3034454003A5070267604 @default.
- W3034454003 hasConcept C111919701 @default.
- W3034454003 hasConcept C11413529 @default.
- W3034454003 hasConcept C119857082 @default.
- W3034454003 hasConcept C127705205 @default.
- W3034454003 hasConcept C13280743 @default.
- W3034454003 hasConcept C134306372 @default.
- W3034454003 hasConcept C146849305 @default.
- W3034454003 hasConcept C153258448 @default.
- W3034454003 hasConcept C154945302 @default.
- W3034454003 hasConcept C177148314 @default.
- W3034454003 hasConcept C182365436 @default.
- W3034454003 hasConcept C185798385 @default.
- W3034454003 hasConcept C205649164 @default.
- W3034454003 hasConcept C206688291 @default.
- W3034454003 hasConcept C33923547 @default.
- W3034454003 hasConcept C41008148 @default.
- W3034454003 hasConcept C49937458 @default.
- W3034454003 hasConcept C50644808 @default.
- W3034454003 hasConcept C51167844 @default.
- W3034454003 hasConceptScore W3034454003C111919701 @default.
- W3034454003 hasConceptScore W3034454003C11413529 @default.
- W3034454003 hasConceptScore W3034454003C119857082 @default.
- W3034454003 hasConceptScore W3034454003C127705205 @default.
- W3034454003 hasConceptScore W3034454003C13280743 @default.
- W3034454003 hasConceptScore W3034454003C134306372 @default.
- W3034454003 hasConceptScore W3034454003C146849305 @default.
- W3034454003 hasConceptScore W3034454003C153258448 @default.
- W3034454003 hasConceptScore W3034454003C154945302 @default.
- W3034454003 hasConceptScore W3034454003C177148314 @default.
- W3034454003 hasConceptScore W3034454003C182365436 @default.
- W3034454003 hasConceptScore W3034454003C185798385 @default.
- W3034454003 hasConceptScore W3034454003C205649164 @default.
- W3034454003 hasConceptScore W3034454003C206688291 @default.
- W3034454003 hasConceptScore W3034454003C33923547 @default.
- W3034454003 hasConceptScore W3034454003C41008148 @default.
- W3034454003 hasConceptScore W3034454003C49937458 @default.
- W3034454003 hasConceptScore W3034454003C50644808 @default.
- W3034454003 hasConceptScore W3034454003C51167844 @default.
- W3034454003 hasOpenAccess W3034454003 @default.
- W3034454003 hasRelatedWork W2107598941 @default.
- W3034454003 hasRelatedWork W2113459411 @default.
- W3034454003 hasRelatedWork W2251405440 @default.
- W3034454003 hasRelatedWork W2769041395 @default.
- W3034454003 hasRelatedWork W2794354164 @default.
- W3034454003 hasRelatedWork W2895604144 @default.
- W3034454003 hasRelatedWork W2903440991 @default.
- W3034454003 hasRelatedWork W2911424454 @default.
- W3034454003 hasRelatedWork W2951122613 @default.
- W3034454003 hasRelatedWork W2961793968 @default.
- W3034454003 hasRelatedWork W2963341956 @default.
- W3034454003 hasRelatedWork W2997918300 @default.
- W3034454003 hasRelatedWork W3008254571 @default.
- W3034454003 hasRelatedWork W3024569256 @default.
- W3034454003 hasRelatedWork W3128889004 @default.
- W3034454003 hasRelatedWork W3157040150 @default.
- W3034454003 hasRelatedWork W9014458 @default.
- W3034454003 hasRelatedWork W2592362528 @default.
- W3034454003 hasRelatedWork W2962778704 @default.
- W3034454003 hasRelatedWork W2962915384 @default.
- W3034454003 hasVolume "1" @default.
- W3034454003 isParatext "false" @default.
- W3034454003 isRetracted "false" @default.
- W3034454003 magId "3034454003" @default.
- W3034454003 workType "article" @default.