Matches in SemOpenAlex for { <https://semopenalex.org/work/W2898894482> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W2898894482 abstract "The developments of deep neural networks (DNN) in recent years have ushered a brand new era of artificial intelligence. DNNs are proved to be excellent in solving very complex problems, e.g., visual recognition and text understanding, to the extent of competing with or even surpassing people. Despite inspiring and encouraging success of DNNs, thorough theoretical analyses still lack to unravel the mystery of their magics. The design of DNN structure is dominated by empirical results in terms of network depth, number of neurons and activations. A few of remarkable works published recently in an attempt to interpret DNNs have established the first glimpses of their internal mechanisms. Nevertheless, research on exploring how DNNs operate is still at the initial stage with plenty of room for refinement. In this paper, we extend precedent research on neural networks with piecewise linear activations (PLNN) concerning linear regions bounds. We present (i) the exact maximal number of linear regions for single layer PLNNs; (ii) a upper bound for multi-layer PLNNs; and (iii) a tighter upper bound for the maximal number of liner regions on rectifier networks. The derived bounds also indirectly explain why deep models are more powerful than shallow counterparts, and how non-linearity of activation functions impacts on expressiveness of networks." @default.
- W2898894482 created "2018-11-09" @default.
- W2898894482 creator A5001420388 @default.
- W2898894482 creator A5048138372 @default.
- W2898894482 date "2018-10-31" @default.
- W2898894482 modified "2023-09-27" @default.
- W2898894482 title "Nearly-tight bounds on linear regions of piecewise linear neural networks." @default.
- W2898894482 cites W1542886316 @default.
- W2898894482 cites W2039415399 @default.
- W2898894482 cites W2089947415 @default.
- W2898894482 cites W2097117768 @default.
- W2898894482 cites W2099471712 @default.
- W2898894482 cites W2103496339 @default.
- W2898894482 cites W2141473882 @default.
- W2898894482 cites W2184045248 @default.
- W2898894482 cites W2194775991 @default.
- W2898894482 cites W2951603627 @default.
- W2898894482 cites W2963950090 @default.
- W2898894482 cites W2963982496 @default.
- W2898894482 cites W3037950864 @default.
- W2898894482 cites W3044803474 @default.
- W2898894482 hasPublicationYear "2018" @default.
- W2898894482 type Work @default.
- W2898894482 sameAs 2898894482 @default.
- W2898894482 citedByCount "3" @default.
- W2898894482 countsByYear W28988944822020 @default.
- W2898894482 countsByYear W28988944822021 @default.
- W2898894482 crossrefType "posted-content" @default.
- W2898894482 hasAuthorship W2898894482A5001420388 @default.
- W2898894482 hasAuthorship W2898894482A5048138372 @default.
- W2898894482 hasConcept C114614502 @default.
- W2898894482 hasConcept C134306372 @default.
- W2898894482 hasConcept C154945302 @default.
- W2898894482 hasConcept C17095337 @default.
- W2898894482 hasConcept C175202392 @default.
- W2898894482 hasConcept C177973122 @default.
- W2898894482 hasConcept C184720557 @default.
- W2898894482 hasConcept C2524010 @default.
- W2898894482 hasConcept C2984842247 @default.
- W2898894482 hasConcept C33923547 @default.
- W2898894482 hasConcept C41008148 @default.
- W2898894482 hasConcept C50100734 @default.
- W2898894482 hasConcept C50644808 @default.
- W2898894482 hasConcept C77553402 @default.
- W2898894482 hasConceptScore W2898894482C114614502 @default.
- W2898894482 hasConceptScore W2898894482C134306372 @default.
- W2898894482 hasConceptScore W2898894482C154945302 @default.
- W2898894482 hasConceptScore W2898894482C17095337 @default.
- W2898894482 hasConceptScore W2898894482C175202392 @default.
- W2898894482 hasConceptScore W2898894482C177973122 @default.
- W2898894482 hasConceptScore W2898894482C184720557 @default.
- W2898894482 hasConceptScore W2898894482C2524010 @default.
- W2898894482 hasConceptScore W2898894482C2984842247 @default.
- W2898894482 hasConceptScore W2898894482C33923547 @default.
- W2898894482 hasConceptScore W2898894482C41008148 @default.
- W2898894482 hasConceptScore W2898894482C50100734 @default.
- W2898894482 hasConceptScore W2898894482C50644808 @default.
- W2898894482 hasConceptScore W2898894482C77553402 @default.
- W2898894482 hasLocation W28988944821 @default.
- W2898894482 hasOpenAccess W2898894482 @default.
- W2898894482 hasPrimaryLocation W28988944821 @default.
- W2898894482 hasRelatedWork W2212676342 @default.
- W2898894482 hasRelatedWork W2785807865 @default.
- W2898894482 hasRelatedWork W2808157231 @default.
- W2898894482 hasRelatedWork W2810699377 @default.
- W2898894482 hasRelatedWork W2914483840 @default.
- W2898894482 hasRelatedWork W2951611096 @default.
- W2898894482 hasRelatedWork W2956024278 @default.
- W2898894482 hasRelatedWork W2964098911 @default.
- W2898894482 hasRelatedWork W2977118166 @default.
- W2898894482 hasRelatedWork W2979730033 @default.
- W2898894482 hasRelatedWork W2995935760 @default.
- W2898894482 hasRelatedWork W2997330246 @default.
- W2898894482 hasRelatedWork W3012683646 @default.
- W2898894482 hasRelatedWork W3042991734 @default.
- W2898894482 hasRelatedWork W3084746729 @default.
- W2898894482 hasRelatedWork W3113190353 @default.
- W2898894482 hasRelatedWork W3164547988 @default.
- W2898894482 hasRelatedWork W3166177065 @default.
- W2898894482 hasRelatedWork W3171401795 @default.
- W2898894482 hasRelatedWork W3200346102 @default.
- W2898894482 isParatext "false" @default.
- W2898894482 isRetracted "false" @default.
- W2898894482 magId "2898894482" @default.
- W2898894482 workType "article" @default.