Matches in SemOpenAlex for { <https://semopenalex.org/work/W2808955470> ?p ?o ?g. }
- W2808955470 abstract "Recurrent neural networks (RNNs) have been drawing much attention with great success in many applications like speech recognition and neural machine translation. Long short-term memory (LSTM) is one of the most popular RNN units in deep learning applications. LSTM transforms the input and the previous hidden states to the next states with the affine transformation, multiplication operations and a nonlinear activation function, which makes a good data representation for a given task. The affine transformation includes rotation and reflection, which change the semantic or syntactic information of dimensions in the hidden states. However, considering that a model interprets the output sequence of LSTM over the whole input sequence, the dimensions of the states need to keep the same type of semantic or syntactic information regardless of the location in the sequence. In this paper, we propose a simple variant of the LSTM unit, persistent recurrent unit (PRU), where each dimension of hidden states keeps persistent information across time, so that the space keeps the same meaning over the whole sequence. In addition, to improve the nonlinear transformation power, we add a feedforward layer in the PRU structure. In the experiment, we evaluate our proposed methods with three different tasks, and the results confirm that our methods have better performance than the conventional LSTM." @default.
- W2808955470 created "2018-06-29" @default.
- W2808955470 creator A5079848178 @default.
- W2808955470 date "2018-06-22" @default.
- W2808955470 modified "2023-09-27" @default.
- W2808955470 title "Persistent Hidden States and Nonlinear Transformation for Long Short-Term Memory" @default.
- W2808955470 cites W1514535095 @default.
- W2808955470 cites W1689711448 @default.
- W2808955470 cites W1771459135 @default.
- W2808955470 cites W179875071 @default.
- W2808955470 cites W1800356822 @default.
- W2808955470 cites W1815076433 @default.
- W2808955470 cites W1915251500 @default.
- W2808955470 cites W1951216520 @default.
- W2808955470 cites W2005708641 @default.
- W2808955470 cites W2132339004 @default.
- W2808955470 cites W2133564696 @default.
- W2808955470 cites W2147568880 @default.
- W2808955470 cites W2172140247 @default.
- W2808955470 cites W2265796482 @default.
- W2808955470 cites W2343954916 @default.
- W2808955470 cites W2525332836 @default.
- W2808955470 cites W2949888546 @default.
- W2808955470 cites W2962784628 @default.
- W2808955470 cites W2962878352 @default.
- W2808955470 cites W2963042606 @default.
- W2808955470 cites W2963277143 @default.
- W2808955470 cites W2964121744 @default.
- W2808955470 cites W3093329015 @default.
- W2808955470 hasPublicationYear "2018" @default.
- W2808955470 type Work @default.
- W2808955470 sameAs 2808955470 @default.
- W2808955470 citedByCount "0" @default.
- W2808955470 crossrefType "posted-content" @default.
- W2808955470 hasAuthorship W2808955470A5079848178 @default.
- W2808955470 hasConcept C104317684 @default.
- W2808955470 hasConcept C108583219 @default.
- W2808955470 hasConcept C114614502 @default.
- W2808955470 hasConcept C121332964 @default.
- W2808955470 hasConcept C147168706 @default.
- W2808955470 hasConcept C154945302 @default.
- W2808955470 hasConcept C158622935 @default.
- W2808955470 hasConcept C17744445 @default.
- W2808955470 hasConcept C185592680 @default.
- W2808955470 hasConcept C199539241 @default.
- W2808955470 hasConcept C202444582 @default.
- W2808955470 hasConcept C204241405 @default.
- W2808955470 hasConcept C2776359362 @default.
- W2808955470 hasConcept C2778112365 @default.
- W2808955470 hasConcept C2780595030 @default.
- W2808955470 hasConcept C28490314 @default.
- W2808955470 hasConcept C33923547 @default.
- W2808955470 hasConcept C40506919 @default.
- W2808955470 hasConcept C41008148 @default.
- W2808955470 hasConcept C50644808 @default.
- W2808955470 hasConcept C54355233 @default.
- W2808955470 hasConcept C55493867 @default.
- W2808955470 hasConcept C62520636 @default.
- W2808955470 hasConcept C86803240 @default.
- W2808955470 hasConcept C92757383 @default.
- W2808955470 hasConcept C94625758 @default.
- W2808955470 hasConceptScore W2808955470C104317684 @default.
- W2808955470 hasConceptScore W2808955470C108583219 @default.
- W2808955470 hasConceptScore W2808955470C114614502 @default.
- W2808955470 hasConceptScore W2808955470C121332964 @default.
- W2808955470 hasConceptScore W2808955470C147168706 @default.
- W2808955470 hasConceptScore W2808955470C154945302 @default.
- W2808955470 hasConceptScore W2808955470C158622935 @default.
- W2808955470 hasConceptScore W2808955470C17744445 @default.
- W2808955470 hasConceptScore W2808955470C185592680 @default.
- W2808955470 hasConceptScore W2808955470C199539241 @default.
- W2808955470 hasConceptScore W2808955470C202444582 @default.
- W2808955470 hasConceptScore W2808955470C204241405 @default.
- W2808955470 hasConceptScore W2808955470C2776359362 @default.
- W2808955470 hasConceptScore W2808955470C2778112365 @default.
- W2808955470 hasConceptScore W2808955470C2780595030 @default.
- W2808955470 hasConceptScore W2808955470C28490314 @default.
- W2808955470 hasConceptScore W2808955470C33923547 @default.
- W2808955470 hasConceptScore W2808955470C40506919 @default.
- W2808955470 hasConceptScore W2808955470C41008148 @default.
- W2808955470 hasConceptScore W2808955470C50644808 @default.
- W2808955470 hasConceptScore W2808955470C54355233 @default.
- W2808955470 hasConceptScore W2808955470C55493867 @default.
- W2808955470 hasConceptScore W2808955470C62520636 @default.
- W2808955470 hasConceptScore W2808955470C86803240 @default.
- W2808955470 hasConceptScore W2808955470C92757383 @default.
- W2808955470 hasConceptScore W2808955470C94625758 @default.
- W2808955470 hasLocation W28089554701 @default.
- W2808955470 hasOpenAccess W2808955470 @default.
- W2808955470 hasPrimaryLocation W28089554701 @default.
- W2808955470 hasRelatedWork W2099052794 @default.
- W2808955470 hasRelatedWork W2586630418 @default.
- W2808955470 hasRelatedWork W2586717708 @default.
- W2808955470 hasRelatedWork W2746131160 @default.
- W2808955470 hasRelatedWork W2790400419 @default.
- W2808955470 hasRelatedWork W2790662531 @default.
- W2808955470 hasRelatedWork W2810658548 @default.
- W2808955470 hasRelatedWork W2898065209 @default.
- W2808955470 hasRelatedWork W2899581658 @default.
- W2808955470 hasRelatedWork W2899603693 @default.