Matches in SemOpenAlex for { <https://semopenalex.org/work/W3125330101> ?p ?o ?g. }
- W3125330101 abstract "In this work, we propose an information theory based framework DeepMI to train deep neural networks (DNN) using Mutual Information (MI). The DeepMI framework is especially targeted but not limited to the learning of real world tasks in an unsupervised manner. The primary motivation behind this work is the insufficiency of traditional loss functions for unsupervised task learning. Moreover, directly using MI for the training purpose is quite challenging to deal because of its unbounded above nature. Hence, we develop an alternative linearized representation of MI as a part of the framework. Contributions of this paper are three fold: i) investigation of MI to train deep neural networks, ii) novel loss function LLMI, and iii) a fuzzy logic based end-to-end differentiable pipeline to integrate DeepMI into deep learning framework. We choose a few unsupervised learning tasks for our experimental study. We demonstrate that L LM I alone provides better gradients to achieve a neural network better performance over the cases when multiple loss functions are used for a given task." @default.
- W3125330101 created "2021-02-01" @default.
- W3125330101 creator A5040749287 @default.
- W3125330101 creator A5065056581 @default.
- W3125330101 date "2021-01-16" @default.
- W3125330101 modified "2023-10-18" @default.
- W3125330101 title "DeepMI: A Mutual Information Based Framework For Unsupervised Deep Learning of Tasks." @default.
- W3125330101 cites W1522301498 @default.
- W3125330101 cites W1536680647 @default.
- W3125330101 cites W1874027545 @default.
- W3125330101 cites W1877570817 @default.
- W3125330101 cites W1901129140 @default.
- W3125330101 cites W1995875735 @default.
- W3125330101 cites W2001613026 @default.
- W3125330101 cites W2034432063 @default.
- W3125330101 cites W2099111195 @default.
- W3125330101 cites W2099471712 @default.
- W3125330101 cites W2108598243 @default.
- W3125330101 cites W2115579991 @default.
- W3125330101 cites W2123649031 @default.
- W3125330101 cites W2123838014 @default.
- W3125330101 cites W2127234432 @default.
- W3125330101 cites W2133665775 @default.
- W3125330101 cites W2134478553 @default.
- W3125330101 cites W2136503713 @default.
- W3125330101 cites W2154053567 @default.
- W3125330101 cites W2194775991 @default.
- W3125330101 cites W2520707372 @default.
- W3125330101 cites W2560023338 @default.
- W3125330101 cites W2783047733 @default.
- W3125330101 cites W2894983388 @default.
- W3125330101 cites W2910510426 @default.
- W3125330101 cites W2949382160 @default.
- W3125330101 cites W2950094539 @default.
- W3125330101 cites W2953106684 @default.
- W3125330101 cites W2962816904 @default.
- W3125330101 cites W2962835968 @default.
- W3125330101 cites W2964314455 @default.
- W3125330101 cites W2968529504 @default.
- W3125330101 cites W2971000934 @default.
- W3125330101 cites W3100901464 @default.
- W3125330101 cites W3106250896 @default.
- W3125330101 cites W603908379 @default.
- W3125330101 hasPublicationYear "2021" @default.
- W3125330101 type Work @default.
- W3125330101 sameAs 3125330101 @default.
- W3125330101 citedByCount "0" @default.
- W3125330101 crossrefType "posted-content" @default.
- W3125330101 hasAuthorship W3125330101A5040749287 @default.
- W3125330101 hasAuthorship W3125330101A5065056581 @default.
- W3125330101 hasConcept C108583219 @default.
- W3125330101 hasConcept C119857082 @default.
- W3125330101 hasConcept C127413603 @default.
- W3125330101 hasConcept C134306372 @default.
- W3125330101 hasConcept C14036430 @default.
- W3125330101 hasConcept C152139883 @default.
- W3125330101 hasConcept C154945302 @default.
- W3125330101 hasConcept C17744445 @default.
- W3125330101 hasConcept C199360897 @default.
- W3125330101 hasConcept C199539241 @default.
- W3125330101 hasConcept C201995342 @default.
- W3125330101 hasConcept C202615002 @default.
- W3125330101 hasConcept C2776359362 @default.
- W3125330101 hasConcept C2780451532 @default.
- W3125330101 hasConcept C2984842247 @default.
- W3125330101 hasConcept C33923547 @default.
- W3125330101 hasConcept C41008148 @default.
- W3125330101 hasConcept C43521106 @default.
- W3125330101 hasConcept C50644808 @default.
- W3125330101 hasConcept C59404180 @default.
- W3125330101 hasConcept C78458016 @default.
- W3125330101 hasConcept C8038995 @default.
- W3125330101 hasConcept C86803240 @default.
- W3125330101 hasConcept C94625758 @default.
- W3125330101 hasConceptScore W3125330101C108583219 @default.
- W3125330101 hasConceptScore W3125330101C119857082 @default.
- W3125330101 hasConceptScore W3125330101C127413603 @default.
- W3125330101 hasConceptScore W3125330101C134306372 @default.
- W3125330101 hasConceptScore W3125330101C14036430 @default.
- W3125330101 hasConceptScore W3125330101C152139883 @default.
- W3125330101 hasConceptScore W3125330101C154945302 @default.
- W3125330101 hasConceptScore W3125330101C17744445 @default.
- W3125330101 hasConceptScore W3125330101C199360897 @default.
- W3125330101 hasConceptScore W3125330101C199539241 @default.
- W3125330101 hasConceptScore W3125330101C201995342 @default.
- W3125330101 hasConceptScore W3125330101C202615002 @default.
- W3125330101 hasConceptScore W3125330101C2776359362 @default.
- W3125330101 hasConceptScore W3125330101C2780451532 @default.
- W3125330101 hasConceptScore W3125330101C2984842247 @default.
- W3125330101 hasConceptScore W3125330101C33923547 @default.
- W3125330101 hasConceptScore W3125330101C41008148 @default.
- W3125330101 hasConceptScore W3125330101C43521106 @default.
- W3125330101 hasConceptScore W3125330101C50644808 @default.
- W3125330101 hasConceptScore W3125330101C59404180 @default.
- W3125330101 hasConceptScore W3125330101C78458016 @default.
- W3125330101 hasConceptScore W3125330101C8038995 @default.
- W3125330101 hasConceptScore W3125330101C86803240 @default.
- W3125330101 hasConceptScore W3125330101C94625758 @default.
- W3125330101 hasLocation W31253301011 @default.
- W3125330101 hasOpenAccess W3125330101 @default.