Matches in SemOpenAlex for { <https://semopenalex.org/work/W2464779257> ?p ?o ?g. }
- W2464779257 endingPage "1434" @default.
- W2464779257 startingPage "1426" @default.
- W2464779257 abstract "The Noisy Non-negative Matrix factorization (NMF) is: given a data matrix A (d × n), find non-negative matrices B, C (d × k, k × n respy.) so that A = BC + N, where N is a noise matrix. Existing polynomial time algorithms with proven error guarantees require each column N.,j to have l1 norm much smaller than ||(BC).,j||1, which could be very restrictive. In important applications of NMF such as Topic Modeling as well as theoretical noise models (eg. Gaussian with high σ), almost every column of N.j violates this condition. We introduce the heavy noise model which only requires the average noise over large subsets of columns to be small. We initiate a study of Noisy NMF under the heavy noise model. We show that our noise model subsumes noise models of theoretical and practical interest (for eg. Gaussian noise of maximum possible σ). We then devise an algorithm TSVDNMF which under certain assumptions on B,C, solves the problem under heavy noise. Our error guarantees match those of previous algorithms. Our running time of O((n + d)2k) is substantially better than the O(n3d) for the previous best. Our assumption on B is weaker than the Separability assumption made by all previous results. We provide empirical justification for our assumptions on C. We also provide the first proof of identifiability (uniqueness of B) for noisy NMF which is not based on separability and does not use hard to check geometric conditions. Our algorithm outperforms earlier polynomial time algorithms both in time and error, particularly in the presence of high noise." @default.
- W2464779257 created "2016-07-22" @default.
- W2464779257 creator A5007922392 @default.
- W2464779257 creator A5029531416 @default.
- W2464779257 creator A5032650028 @default.
- W2464779257 creator A5047569978 @default.
- W2464779257 date "2016-06-19" @default.
- W2464779257 modified "2023-10-16" @default.
- W2464779257 title "Non-negative matrix factorization under heavy noise" @default.
- W2464779257 cites W1524090374 @default.
- W2464779257 cites W1532325895 @default.
- W2464779257 cites W1854811422 @default.
- W2464779257 cites W1882879704 @default.
- W2464779257 cites W1902027874 @default.
- W2464779257 cites W1982755765 @default.
- W2464779257 cites W2048687561 @default.
- W2464779257 cites W2070135644 @default.
- W2464779257 cites W2105503503 @default.
- W2464779257 cites W2109685970 @default.
- W2464779257 cites W2111604514 @default.
- W2464779257 cites W2114760097 @default.
- W2464779257 cites W2117986441 @default.
- W2464779257 cites W2125118959 @default.
- W2464779257 cites W2125477383 @default.
- W2464779257 cites W2140318696 @default.
- W2464779257 cites W2951734015 @default.
- W2464779257 cites W2989661724 @default.
- W2464779257 cites W3099035708 @default.
- W2464779257 hasPublicationYear "2016" @default.
- W2464779257 type Work @default.
- W2464779257 sameAs 2464779257 @default.
- W2464779257 citedByCount "4" @default.
- W2464779257 countsByYear W24647792572016 @default.
- W2464779257 countsByYear W24647792572017 @default.
- W2464779257 countsByYear W24647792572019 @default.
- W2464779257 countsByYear W24647792572021 @default.
- W2464779257 crossrefType "proceedings-article" @default.
- W2464779257 hasAuthorship W2464779257A5007922392 @default.
- W2464779257 hasAuthorship W2464779257A5029531416 @default.
- W2464779257 hasAuthorship W2464779257A5032650028 @default.
- W2464779257 hasAuthorship W2464779257A5047569978 @default.
- W2464779257 hasConcept C105795698 @default.
- W2464779257 hasConcept C106487976 @default.
- W2464779257 hasConcept C11413529 @default.
- W2464779257 hasConcept C115961682 @default.
- W2464779257 hasConcept C121332964 @default.
- W2464779257 hasConcept C122770356 @default.
- W2464779257 hasConcept C134306372 @default.
- W2464779257 hasConcept C152671427 @default.
- W2464779257 hasConcept C154945302 @default.
- W2464779257 hasConcept C158693339 @default.
- W2464779257 hasConcept C159985019 @default.
- W2464779257 hasConcept C187834632 @default.
- W2464779257 hasConcept C192562407 @default.
- W2464779257 hasConcept C2777021972 @default.
- W2464779257 hasConcept C28826006 @default.
- W2464779257 hasConcept C33923547 @default.
- W2464779257 hasConcept C41008148 @default.
- W2464779257 hasConcept C4199805 @default.
- W2464779257 hasConcept C42355184 @default.
- W2464779257 hasConcept C62520636 @default.
- W2464779257 hasConcept C90119067 @default.
- W2464779257 hasConcept C99498987 @default.
- W2464779257 hasConceptScore W2464779257C105795698 @default.
- W2464779257 hasConceptScore W2464779257C106487976 @default.
- W2464779257 hasConceptScore W2464779257C11413529 @default.
- W2464779257 hasConceptScore W2464779257C115961682 @default.
- W2464779257 hasConceptScore W2464779257C121332964 @default.
- W2464779257 hasConceptScore W2464779257C122770356 @default.
- W2464779257 hasConceptScore W2464779257C134306372 @default.
- W2464779257 hasConceptScore W2464779257C152671427 @default.
- W2464779257 hasConceptScore W2464779257C154945302 @default.
- W2464779257 hasConceptScore W2464779257C158693339 @default.
- W2464779257 hasConceptScore W2464779257C159985019 @default.
- W2464779257 hasConceptScore W2464779257C187834632 @default.
- W2464779257 hasConceptScore W2464779257C192562407 @default.
- W2464779257 hasConceptScore W2464779257C2777021972 @default.
- W2464779257 hasConceptScore W2464779257C28826006 @default.
- W2464779257 hasConceptScore W2464779257C33923547 @default.
- W2464779257 hasConceptScore W2464779257C41008148 @default.
- W2464779257 hasConceptScore W2464779257C4199805 @default.
- W2464779257 hasConceptScore W2464779257C42355184 @default.
- W2464779257 hasConceptScore W2464779257C62520636 @default.
- W2464779257 hasConceptScore W2464779257C90119067 @default.
- W2464779257 hasConceptScore W2464779257C99498987 @default.
- W2464779257 hasLocation W24647792571 @default.
- W2464779257 hasOpenAccess W2464779257 @default.
- W2464779257 hasPrimaryLocation W24647792571 @default.
- W2464779257 hasRelatedWork W1902027874 @default.
- W2464779257 hasRelatedWork W2166581276 @default.
- W2464779257 hasRelatedWork W2340619382 @default.
- W2464779257 hasRelatedWork W2800096518 @default.
- W2464779257 hasRelatedWork W2951620362 @default.
- W2464779257 hasRelatedWork W2951870128 @default.
- W2464779257 hasRelatedWork W2953167782 @default.
- W2464779257 hasRelatedWork W2954185869 @default.
- W2464779257 hasRelatedWork W2979328751 @default.
- W2464779257 hasRelatedWork W3013109894 @default.