FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2950858170> ?p ?o ?g. }

• W2950858170 abstract "Optimization of machine learning models is commonly performed through stochastic gradient updates on randomly ordered training examples. This practice means that sub-epochs comprise of independent random samples of the training data that may not preserve informative structure present in the full data. We hypothesize that the training can be more effective with {em self-similar} arrangements that potentially allow each epoch to provide benefits of multiple ones. We study this for matrix factorization -- the common task of learning metric embeddings of entities such as queries, videos, or words from example pairwise associations. We construct arrangements that preserve the weighted Jaccard similarities of rows and columns and experimentally observe training acceleration of 3%-37% on synthetic and recommendation datasets. Principled arrangements of training examples emerge as a novel and potentially powerful enhancement to SGD that merits further exploration." @default.
• W2950858170 created "2019-06-27" @default.
• W2950858170 creator A5026385549 @default.
• W2950858170 creator A5065128060 @default.
• W2950858170 creator A5089144342 @default.
• W2950858170 creator A5089579452 @default.
• W2950858170 date "2018-03-14" @default.
• W2950858170 modified "2023-09-27" @default.
• W2950858170 title "Self-Similar Epochs: Value in Arrangement" @default.
• W2950858170 cites W1577337550 @default.
• W2950858170 cites W1965996575 @default.
• W2950858170 cites W1978394996 @default.
• W2950858170 cites W1980147176 @default.
• W2950858170 cites W1985123706 @default.
• W2950858170 cites W1994389483 @default.
• W2950858170 cites W2018456433 @default.
• W2950858170 cites W2024128809 @default.
• W2950858170 cites W2047620357 @default.
• W2950858170 cites W2054141820 @default.
• W2950858170 cites W2072773380 @default.
• W2950858170 cites W2077499269 @default.
• W2950858170 cites W2092236286 @default.
• W2950858170 cites W2099866409 @default.
• W2950858170 cites W2101409192 @default.
• W2950858170 cites W2108598243 @default.
• W2950858170 cites W2127589108 @default.
• W2950858170 cites W2146502635 @default.
• W2950858170 cites W2148885851 @default.
• W2950858170 cites W2153579005 @default.
• W2950858170 cites W2157364932 @default.
• W2950858170 cites W2169901824 @default.
• W2950858170 cites W2177410802 @default.
• W2950858170 cites W2250539671 @default.
• W2950858170 cites W2732240984 @default.
• W2950858170 cites W3104097132 @default.
• W2950858170 cites W3193262203 @default.
• W2950858170 cites W195445992 @default.
• W2950858170 hasPublicationYear "2018" @default.
• W2950858170 type Work @default.
• W2950858170 sameAs 2950858170 @default.
• W2950858170 citedByCount "0" @default.
• W2950858170 crossrefType "posted-content" @default.
• W2950858170 hasAuthorship W2950858170A5026385549 @default.
• W2950858170 hasAuthorship W2950858170A5065128060 @default.
• W2950858170 hasAuthorship W2950858170A5089144342 @default.
• W2950858170 hasAuthorship W2950858170A5089579452 @default.
• W2950858170 hasConcept C119857082 @default.
• W2950858170 hasConcept C121332964 @default.
• W2950858170 hasConcept C135598885 @default.
• W2950858170 hasConcept C153180895 @default.
• W2950858170 hasConcept C154945302 @default.
• W2950858170 hasConcept C158693339 @default.
• W2950858170 hasConcept C162324750 @default.
• W2950858170 hasConcept C176217482 @default.
• W2950858170 hasConcept C184898388 @default.
• W2950858170 hasConcept C187736073 @default.
• W2950858170 hasConcept C199360897 @default.
• W2950858170 hasConcept C203519979 @default.
• W2950858170 hasConcept C21547014 @default.
• W2950858170 hasConcept C2780451532 @default.
• W2950858170 hasConcept C2780801425 @default.
• W2950858170 hasConcept C41008148 @default.
• W2950858170 hasConcept C42355184 @default.
• W2950858170 hasConcept C62520636 @default.
• W2950858170 hasConcept C77088390 @default.
• W2950858170 hasConceptScore W2950858170C119857082 @default.
• W2950858170 hasConceptScore W2950858170C121332964 @default.
• W2950858170 hasConceptScore W2950858170C135598885 @default.
• W2950858170 hasConceptScore W2950858170C153180895 @default.
• W2950858170 hasConceptScore W2950858170C154945302 @default.
• W2950858170 hasConceptScore W2950858170C158693339 @default.
• W2950858170 hasConceptScore W2950858170C162324750 @default.
• W2950858170 hasConceptScore W2950858170C176217482 @default.
• W2950858170 hasConceptScore W2950858170C184898388 @default.
• W2950858170 hasConceptScore W2950858170C187736073 @default.
• W2950858170 hasConceptScore W2950858170C199360897 @default.
• W2950858170 hasConceptScore W2950858170C203519979 @default.
• W2950858170 hasConceptScore W2950858170C21547014 @default.
• W2950858170 hasConceptScore W2950858170C2780451532 @default.
• W2950858170 hasConceptScore W2950858170C2780801425 @default.
• W2950858170 hasConceptScore W2950858170C41008148 @default.
• W2950858170 hasConceptScore W2950858170C42355184 @default.
• W2950858170 hasConceptScore W2950858170C62520636 @default.
• W2950858170 hasConceptScore W2950858170C77088390 @default.
• W2950858170 hasLocation W29508581701 @default.
• W2950858170 hasOpenAccess W2950858170 @default.
• W2950858170 hasPrimaryLocation W29508581701 @default.
• W2950858170 hasRelatedWork W1495279923 @default.
• W2950858170 hasRelatedWork W2042714649 @default.
• W2950858170 hasRelatedWork W2275114322 @default.
• W2950858170 hasRelatedWork W2299623326 @default.
• W2950858170 hasRelatedWork W2530382149 @default.
• W2950858170 hasRelatedWork W2600135312 @default.
• W2950858170 hasRelatedWork W2604829508 @default.
• W2950858170 hasRelatedWork W2753880617 @default.
• W2950858170 hasRelatedWork W2900528039 @default.
• W2950858170 hasRelatedWork W2963877826 @default.
• W2950858170 hasRelatedWork W2990726673 @default.
• W2950858170 hasRelatedWork W3004169498 @default.
• W2950858170 hasRelatedWork W3005343217 @default.

• next