FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2726793048> ?p ?o ?g. }

• W2726793048 abstract "We study the efficient learnability of geometric concept classes - specifically, low-degree polynomial threshold functions (PTFs) and intersections of halfspaces - when a fraction of the data is adversarially corrupted. We give the first polynomial-time PAC learning algorithms for these concept classes with dimension-independent error guarantees in the presence of nasty noise under the Gaussian distribution. In the nasty noise model, an omniscient adversary can arbitrarily corrupt a small fraction of both the unlabeled data points and their labels. This model generalizes well-studied noise models, including the malicious noise model and the agnostic (adversarial label noise) model. Prior to our work, the only concept class for which efficient malicious learning algorithms were known was the class of origin-centered halfspaces. Specifically, our robust learning algorithm for low-degree PTFs succeeds under a number of tame distributions -- including the Gaussian distribution and, more generally, any log-concave distribution with (approximately) known low-degree moments. For LTFs under the Gaussian distribution, we give a polynomial-time algorithm that achieves error $O(epsilon)$, where $epsilon$ is the noise rate. At the core of our PAC learning results is an efficient algorithm to approximate the low-degree Chow-parameters of any bounded function in the presence of nasty noise. To achieve this, we employ an iterative spectral method for outlier detection and removal, inspired by recent work in robust unsupervised learning. Our aforementioned algorithm succeeds for a range of distributions satisfying mild concentration bounds and moment assumptions. The correctness of our robust learning algorithm for intersections of halfspaces makes essential use of a novel robust inverse independence lemma that may be of broader interest." @default.
• W2726793048 created "2017-07-14" @default.
• W2726793048 creator A5035962853 @default.
• W2726793048 creator A5054966819 @default.
• W2726793048 creator A5083699810 @default.
• W2726793048 date "2017-07-05" @default.
• W2726793048 modified "2023-09-27" @default.
• W2726793048 title "Learning Geometric Concepts with Nasty Noise" @default.
• W2726793048 cites W1486564402 @default.
• W2726793048 cites W1528486370 @default.
• W2726793048 cites W1585566614 @default.
• W2726793048 cites W1609370892 @default.
• W2726793048 cites W1965187358 @default.
• W2726793048 cites W1968998685 @default.
• W2726793048 cites W1997142544 @default.
• W2726793048 cites W2019120673 @default.
• W2726793048 cites W2019363670 @default.
• W2726793048 cites W2022945027 @default.
• W2726793048 cites W2046242327 @default.
• W2726793048 cites W2071833075 @default.
• W2726793048 cites W2084544490 @default.
• W2726793048 cites W2086789740 @default.
• W2726793048 cites W2097575812 @default.
• W2726793048 cites W2106458073 @default.
• W2726793048 cites W2122170585 @default.
• W2726793048 cites W2140125566 @default.
• W2726793048 cites W2163332813 @default.
• W2726793048 cites W2180443860 @default.
• W2726793048 cites W2201177080 @default.
• W2726793048 cites W2256202 @default.
• W2726793048 cites W2545696445 @default.
• W2726793048 cites W2567351292 @default.
• W2726793048 cites W2592318711 @default.
• W2726793048 cites W3021023808 @default.
• W2726793048 cites W3100065552 @default.
• W2726793048 hasPublicationYear "2017" @default.
• W2726793048 type Work @default.
• W2726793048 sameAs 2726793048 @default.
• W2726793048 citedByCount "6" @default.
• W2726793048 countsByYear W27267930482017 @default.
• W2726793048 countsByYear W27267930482018 @default.
• W2726793048 countsByYear W27267930482019 @default.
• W2726793048 crossrefType "posted-content" @default.
• W2726793048 hasAuthorship W2726793048A5035962853 @default.
• W2726793048 hasAuthorship W2726793048A5054966819 @default.
• W2726793048 hasAuthorship W2726793048A5083699810 @default.
• W2726793048 hasConcept C11413529 @default.
• W2726793048 hasConcept C114614502 @default.
• W2726793048 hasConcept C115961682 @default.
• W2726793048 hasConcept C121332964 @default.
• W2726793048 hasConcept C134306372 @default.
• W2726793048 hasConcept C149629883 @default.
• W2726793048 hasConcept C154945302 @default.
• W2726793048 hasConcept C163716315 @default.
• W2726793048 hasConcept C178790620 @default.
• W2726793048 hasConcept C185592680 @default.
• W2726793048 hasConcept C24890656 @default.
• W2726793048 hasConcept C2775997480 @default.
• W2726793048 hasConcept C2777723229 @default.
• W2726793048 hasConcept C33676613 @default.
• W2726793048 hasConcept C33923547 @default.
• W2726793048 hasConcept C34388435 @default.
• W2726793048 hasConcept C41008148 @default.
• W2726793048 hasConcept C55439883 @default.
• W2726793048 hasConcept C62520636 @default.
• W2726793048 hasConcept C79337645 @default.
• W2726793048 hasConcept C90119067 @default.
• W2726793048 hasConcept C99498987 @default.
• W2726793048 hasConceptScore W2726793048C11413529 @default.
• W2726793048 hasConceptScore W2726793048C114614502 @default.
• W2726793048 hasConceptScore W2726793048C115961682 @default.
• W2726793048 hasConceptScore W2726793048C121332964 @default.
• W2726793048 hasConceptScore W2726793048C134306372 @default.
• W2726793048 hasConceptScore W2726793048C149629883 @default.
• W2726793048 hasConceptScore W2726793048C154945302 @default.
• W2726793048 hasConceptScore W2726793048C163716315 @default.
• W2726793048 hasConceptScore W2726793048C178790620 @default.
• W2726793048 hasConceptScore W2726793048C185592680 @default.
• W2726793048 hasConceptScore W2726793048C24890656 @default.
• W2726793048 hasConceptScore W2726793048C2775997480 @default.
• W2726793048 hasConceptScore W2726793048C2777723229 @default.
• W2726793048 hasConceptScore W2726793048C33676613 @default.
• W2726793048 hasConceptScore W2726793048C33923547 @default.
• W2726793048 hasConceptScore W2726793048C34388435 @default.
• W2726793048 hasConceptScore W2726793048C41008148 @default.
• W2726793048 hasConceptScore W2726793048C55439883 @default.
• W2726793048 hasConceptScore W2726793048C62520636 @default.
• W2726793048 hasConceptScore W2726793048C79337645 @default.
• W2726793048 hasConceptScore W2726793048C90119067 @default.
• W2726793048 hasConceptScore W2726793048C99498987 @default.
• W2726793048 hasLocation W27267930481 @default.
• W2726793048 hasOpenAccess W2726793048 @default.
• W2726793048 hasPrimaryLocation W27267930481 @default.
• W2726793048 hasRelatedWork W2072782484 @default.
• W2726793048 hasRelatedWork W2089361539 @default.
• W2726793048 hasRelatedWork W2168002876 @default.
• W2726793048 hasRelatedWork W2201177080 @default.
• W2726793048 hasRelatedWork W2460885807 @default.
• W2726793048 hasRelatedWork W2554864439 @default.
• W2726793048 hasRelatedWork W2592318711 @default.

• next