Matches in SemOpenAlex for { <https://semopenalex.org/work/W2950135596> ?p ?o ?g. }
- W2950135596 abstract "The development of cluster computing frameworks has allowed practitioners to scale out various statistical estimation and machine learning algorithms with minimal programming effort. This is especially true for machine learning problems whose objective function is nicely separable across individual data points, such as classification and regression. In contrast, statistical learning tasks involving pairs (or more generally tuples) of data points - such as metric learning, clustering or ranking do not lend themselves as easily to data-parallelism and in-memory computing. In this paper, we investigate how to balance between statistical performance and computational efficiency in such distributed tuplewise statistical problems. We first propose a simple strategy based on occasionally repartitioning data across workers between parallel computation stages, where the number of repartitioning steps rules the trade-off between accuracy and runtime. We then present some theoretical results highlighting the benefits brought by the proposed method in terms of variance reduction, and extend our results to design distributed stochastic gradient descent algorithms for tuplewise empirical risk minimization. Our results are supported by numerical experiments in pairwise statistical estimation and learning on synthetic and real-world datasets." @default.
- W2950135596 created "2019-06-27" @default.
- W2950135596 creator A5014504793 @default.
- W2950135596 creator A5019971124 @default.
- W2950135596 creator A5046622648 @default.
- W2950135596 creator A5088738795 @default.
- W2950135596 creator A5090342394 @default.
- W2950135596 date "2019-06-21" @default.
- W2950135596 modified "2023-10-18" @default.
- W2950135596 title "Trade-offs in Large-Scale Distributed Tuplewise Estimation and Learning" @default.
- W2950135596 cites W1544839292 @default.
- W2950135596 cites W2009537245 @default.
- W2950135596 cites W2013151705 @default.
- W2950135596 cites W2023049771 @default.
- W2950135596 cites W2023449938 @default.
- W2950135596 cites W2027947344 @default.
- W2950135596 cites W2065230098 @default.
- W2950135596 cites W2113651538 @default.
- W2950135596 cites W2164278908 @default.
- W2950135596 cites W2173213060 @default.
- W2950135596 cites W2185959272 @default.
- W2950135596 cites W2189465200 @default.
- W2950135596 cites W2566979091 @default.
- W2950135596 cites W2579247884 @default.
- W2950135596 cites W2798167046 @default.
- W2950135596 cites W2962710568 @default.
- W2950135596 cites W2962741697 @default.
- W2950135596 cites W2962897077 @default.
- W2950135596 cites W2963075699 @default.
- W2950135596 cites W2963326510 @default.
- W2950135596 cites W2963405668 @default.
- W2950135596 cites W2963834123 @default.
- W2950135596 hasPublicationYear "2019" @default.
- W2950135596 type Work @default.
- W2950135596 sameAs 2950135596 @default.
- W2950135596 citedByCount "0" @default.
- W2950135596 crossrefType "posted-content" @default.
- W2950135596 hasAuthorship W2950135596A5014504793 @default.
- W2950135596 hasAuthorship W2950135596A5019971124 @default.
- W2950135596 hasAuthorship W2950135596A5046622648 @default.
- W2950135596 hasAuthorship W2950135596A5088738795 @default.
- W2950135596 hasAuthorship W2950135596A5090342394 @default.
- W2950135596 hasConcept C107321475 @default.
- W2950135596 hasConcept C118615104 @default.
- W2950135596 hasConcept C118930307 @default.
- W2950135596 hasConcept C119857082 @default.
- W2950135596 hasConcept C124101348 @default.
- W2950135596 hasConcept C126838900 @default.
- W2950135596 hasConcept C154945302 @default.
- W2950135596 hasConcept C162324750 @default.
- W2950135596 hasConcept C176217482 @default.
- W2950135596 hasConcept C183115368 @default.
- W2950135596 hasConcept C184898388 @default.
- W2950135596 hasConcept C189430467 @default.
- W2950135596 hasConcept C206688291 @default.
- W2950135596 hasConcept C21547014 @default.
- W2950135596 hasConcept C33923547 @default.
- W2950135596 hasConcept C41008148 @default.
- W2950135596 hasConcept C50644808 @default.
- W2950135596 hasConcept C71924100 @default.
- W2950135596 hasConcept C73555534 @default.
- W2950135596 hasConceptScore W2950135596C107321475 @default.
- W2950135596 hasConceptScore W2950135596C118615104 @default.
- W2950135596 hasConceptScore W2950135596C118930307 @default.
- W2950135596 hasConceptScore W2950135596C119857082 @default.
- W2950135596 hasConceptScore W2950135596C124101348 @default.
- W2950135596 hasConceptScore W2950135596C126838900 @default.
- W2950135596 hasConceptScore W2950135596C154945302 @default.
- W2950135596 hasConceptScore W2950135596C162324750 @default.
- W2950135596 hasConceptScore W2950135596C176217482 @default.
- W2950135596 hasConceptScore W2950135596C183115368 @default.
- W2950135596 hasConceptScore W2950135596C184898388 @default.
- W2950135596 hasConceptScore W2950135596C189430467 @default.
- W2950135596 hasConceptScore W2950135596C206688291 @default.
- W2950135596 hasConceptScore W2950135596C21547014 @default.
- W2950135596 hasConceptScore W2950135596C33923547 @default.
- W2950135596 hasConceptScore W2950135596C41008148 @default.
- W2950135596 hasConceptScore W2950135596C50644808 @default.
- W2950135596 hasConceptScore W2950135596C71924100 @default.
- W2950135596 hasConceptScore W2950135596C73555534 @default.
- W2950135596 hasLocation W29501355961 @default.
- W2950135596 hasOpenAccess W2950135596 @default.
- W2950135596 hasPrimaryLocation W29501355961 @default.
- W2950135596 hasRelatedWork W1540644317 @default.
- W2950135596 hasRelatedWork W184546935 @default.
- W2950135596 hasRelatedWork W2098003004 @default.
- W2950135596 hasRelatedWork W2114109296 @default.
- W2950135596 hasRelatedWork W2137481367 @default.
- W2950135596 hasRelatedWork W2139059214 @default.
- W2950135596 hasRelatedWork W2157998899 @default.
- W2950135596 hasRelatedWork W2285149252 @default.
- W2950135596 hasRelatedWork W2305870132 @default.
- W2950135596 hasRelatedWork W2607462057 @default.
- W2950135596 hasRelatedWork W2774644914 @default.
- W2950135596 hasRelatedWork W2942014936 @default.
- W2950135596 hasRelatedWork W2963478242 @default.
- W2950135596 hasRelatedWork W2963923417 @default.
- W2950135596 hasRelatedWork W2972449768 @default.
- W2950135596 hasRelatedWork W3080700717 @default.
- W2950135596 hasRelatedWork W3109542001 @default.