Matches in SemOpenAlex for { <https://semopenalex.org/work/W2189388227> ?p ?o ?g. }
- W2189388227 abstract "Deep generative neural networks such as the Deep Belief Network and Deep Boltzmann Machines have been used successfully to model high dimensional visual data. However, they are not robust to common variations such as occlusion and random noise. In this thesis, we explore two strategies for improving the robustness of DBNs. First, we show that a DBN with sparse connections in the first layer is more robust to variations that are not in the training set. Second, we develop a probabilistic denoising algorithm to determine a subset of the hidden layer nodes to unclamp. We show that this can be applied to any feedforward network classifier with localized first layer connections. By utilizing the already available generative model for denoising prior to recognition, we show significantly better performance over the standard DBN implementations for various sources of noise on the standard and Variations MNIST databases." @default.
- W2189388227 created "2016-06-24" @default.
- W2189388227 creator A5034956679 @default.
- W2189388227 date "2010-08-25" @default.
- W2189388227 modified "2023-09-27" @default.
- W2189388227 title "Robust Visual Recognition Using Multilayer Generative Neural Networks" @default.
- W2189388227 cites W1480376833 @default.
- W2189388227 cites W1513873506 @default.
- W2189388227 cites W1526741802 @default.
- W2189388227 cites W1543983856 @default.
- W2189388227 cites W1571295451 @default.
- W2189388227 cites W1624854622 @default.
- W2189388227 cites W1675937554 @default.
- W2189388227 cites W177847060 @default.
- W2189388227 cites W1802356529 @default.
- W2189388227 cites W1813659000 @default.
- W2189388227 cites W1860099275 @default.
- W2189388227 cites W189596042 @default.
- W2189388227 cites W1928278792 @default.
- W2189388227 cites W1963079441 @default.
- W2189388227 cites W1971017968 @default.
- W2189388227 cites W1990838964 @default.
- W2189388227 cites W1993845689 @default.
- W2189388227 cites W1994197834 @default.
- W2189388227 cites W1994616650 @default.
- W2189388227 cites W2013239224 @default.
- W2189388227 cites W2032003058 @default.
- W2189388227 cites W2038735287 @default.
- W2189388227 cites W2042492924 @default.
- W2189388227 cites W2050271480 @default.
- W2189388227 cites W205159212 @default.
- W2189388227 cites W2073783260 @default.
- W2189388227 cites W2074376560 @default.
- W2189388227 cites W2075187489 @default.
- W2189388227 cites W2082308025 @default.
- W2189388227 cites W2083380015 @default.
- W2189388227 cites W2084622166 @default.
- W2189388227 cites W2096192494 @default.
- W2189388227 cites W2100495367 @default.
- W2189388227 cites W2101307791 @default.
- W2189388227 cites W2103212315 @default.
- W2189388227 cites W2104978738 @default.
- W2189388227 cites W2105464770 @default.
- W2189388227 cites W2108665656 @default.
- W2189388227 cites W2110052721 @default.
- W2189388227 cites W2110798204 @default.
- W2189388227 cites W2112796928 @default.
- W2189388227 cites W2113606819 @default.
- W2189388227 cites W2116064496 @default.
- W2189388227 cites W2116825644 @default.
- W2189388227 cites W2124914669 @default.
- W2189388227 cites W2125663122 @default.
- W2189388227 cites W2126060993 @default.
- W2189388227 cites W2128354257 @default.
- W2189388227 cites W2128958079 @default.
- W2189388227 cites W2130325614 @default.
- W2189388227 cites W2131470433 @default.
- W2189388227 cites W2131686571 @default.
- W2189388227 cites W2133257461 @default.
- W2189388227 cites W2134653808 @default.
- W2189388227 cites W2136163184 @default.
- W2189388227 cites W2136922672 @default.
- W2189388227 cites W2138448681 @default.
- W2189388227 cites W2138621090 @default.
- W2189388227 cites W2143797877 @default.
- W2189388227 cites W2145889472 @default.
- W2189388227 cites W2149194912 @default.
- W2189388227 cites W2149369122 @default.
- W2189388227 cites W2150114059 @default.
- W2189388227 cites W2151103935 @default.
- W2189388227 cites W2152826865 @default.
- W2189388227 cites W2158164339 @default.
- W2189388227 cites W2159737176 @default.
- W2189388227 cites W2161000554 @default.
- W2189388227 cites W2161893161 @default.
- W2189388227 cites W2165720259 @default.
- W2189388227 cites W2167608136 @default.
- W2189388227 cites W2169953019 @default.
- W2189388227 cites W2171145682 @default.
- W2189388227 cites W2171262059 @default.
- W2189388227 cites W2172174689 @default.
- W2189388227 cites W2546302380 @default.
- W2189388227 cites W2613634265 @default.
- W2189388227 cites W2616465717 @default.
- W2189388227 cites W3214102110 @default.
- W2189388227 cites W2170302354 @default.
- W2189388227 hasPublicationYear "2010" @default.
- W2189388227 type Work @default.
- W2189388227 sameAs 2189388227 @default.
- W2189388227 citedByCount "0" @default.
- W2189388227 crossrefType "dissertation" @default.
- W2189388227 hasAuthorship W2189388227A5034956679 @default.
- W2189388227 hasConcept C104317684 @default.
- W2189388227 hasConcept C108583219 @default.
- W2189388227 hasConcept C119857082 @default.
- W2189388227 hasConcept C153180895 @default.
- W2189388227 hasConcept C154945302 @default.
- W2189388227 hasConcept C167966045 @default.
- W2189388227 hasConcept C185592680 @default.
- W2189388227 hasConcept C190502265 @default.