Matches in SemOpenAlex for { <https://semopenalex.org/work/W3129945678> ?p ?o ?g. }
- W3129945678 abstract "The study of deep neural networks (DNNs) in the infinite-width limit, via the so-called neural tangent kernel (NTK) approach, has provided new insights into the dynamics of learning, the capacity of generalization, and the impact of initialization. One key DNN architecture remains to be kernelized, namely, the recurrent neural network (RNN). In this paper we introduce and study the Recurrent Neural Tangent Kernel (RNTK), which provides new insights into the behavior of overparametrized RNNs, including how different time steps are weighted by the RNTK to form the output under different initialization parameters and nonlinearity choices, and how inputs of different lengths are treated. The ability to compare inputs of different length is a key property of the RNTK that should greatly benefit practitioners. We demonstrate via a synthetic and 53 real-world data experiments that the RNTK offers significant performance gains over other kernels, including standard NTKs, across a wide array of data sets." @default.
- W3129945678 created "2021-03-01" @default.
- W3129945678 creator A5016499384 @default.
- W3129945678 creator A5018162534 @default.
- W3129945678 creator A5047293370 @default.
- W3129945678 creator A5072713767 @default.
- W3129945678 date "2021-05-03" @default.
- W3129945678 modified "2023-09-24" @default.
- W3129945678 title "The Recurrent Neural Tangent Kernel" @default.
- W3129945678 cites W1567512734 @default.
- W3129945678 cites W1589901016 @default.
- W3129945678 cites W1800356822 @default.
- W3129945678 cites W2064675550 @default.
- W3129945678 cites W2107878631 @default.
- W3129945678 cites W2110485445 @default.
- W3129945678 cites W2153635508 @default.
- W3129945678 cites W2157331557 @default.
- W3129945678 cites W2167608136 @default.
- W3129945678 cites W2474920236 @default.
- W3129945678 cites W2581719241 @default.
- W3129945678 cites W2809090039 @default.
- W3129945678 cites W2900959181 @default.
- W3129945678 cites W2906432502 @default.
- W3129945678 cites W2910655610 @default.
- W3129945678 cites W2910905005 @default.
- W3129945678 cites W2911867426 @default.
- W3129945678 cites W2912260645 @default.
- W3129945678 cites W2913473169 @default.
- W3129945678 cites W2962698540 @default.
- W3129945678 cites W2963042606 @default.
- W3129945678 cites W2963323437 @default.
- W3129945678 cites W2963518130 @default.
- W3129945678 cites W2963886912 @default.
- W3129945678 cites W2964052793 @default.
- W3129945678 cites W2964313743 @default.
- W3129945678 cites W2970330753 @default.
- W3129945678 cites W2970474218 @default.
- W3129945678 cites W2971043187 @default.
- W3129945678 cites W2981549002 @default.
- W3129945678 cites W2991290085 @default.
- W3129945678 cites W2994872659 @default.
- W3129945678 cites W2995097486 @default.
- W3129945678 cites W3005679856 @default.
- W3129945678 cites W3038074040 @default.
- W3129945678 cites W3088274487 @default.
- W3129945678 cites W3101069636 @default.
- W3129945678 hasPublicationYear "2021" @default.
- W3129945678 type Work @default.
- W3129945678 sameAs 3129945678 @default.
- W3129945678 citedByCount "14" @default.
- W3129945678 countsByYear W31299456782020 @default.
- W3129945678 countsByYear W31299456782021 @default.
- W3129945678 crossrefType "proceedings-article" @default.
- W3129945678 hasAuthorship W3129945678A5016499384 @default.
- W3129945678 hasAuthorship W3129945678A5018162534 @default.
- W3129945678 hasAuthorship W3129945678A5047293370 @default.
- W3129945678 hasAuthorship W3129945678A5072713767 @default.
- W3129945678 hasConcept C11413529 @default.
- W3129945678 hasConcept C114466953 @default.
- W3129945678 hasConcept C118615104 @default.
- W3129945678 hasConcept C119857082 @default.
- W3129945678 hasConcept C121332964 @default.
- W3129945678 hasConcept C122280245 @default.
- W3129945678 hasConcept C12267149 @default.
- W3129945678 hasConcept C134306372 @default.
- W3129945678 hasConcept C138187205 @default.
- W3129945678 hasConcept C147168706 @default.
- W3129945678 hasConcept C151201525 @default.
- W3129945678 hasConcept C153180895 @default.
- W3129945678 hasConcept C154945302 @default.
- W3129945678 hasConcept C158622935 @default.
- W3129945678 hasConcept C177148314 @default.
- W3129945678 hasConcept C199360897 @default.
- W3129945678 hasConcept C2524010 @default.
- W3129945678 hasConcept C26517878 @default.
- W3129945678 hasConcept C33923547 @default.
- W3129945678 hasConcept C38652104 @default.
- W3129945678 hasConcept C41008148 @default.
- W3129945678 hasConcept C50644808 @default.
- W3129945678 hasConcept C62520636 @default.
- W3129945678 hasConcept C74193536 @default.
- W3129945678 hasConceptScore W3129945678C11413529 @default.
- W3129945678 hasConceptScore W3129945678C114466953 @default.
- W3129945678 hasConceptScore W3129945678C118615104 @default.
- W3129945678 hasConceptScore W3129945678C119857082 @default.
- W3129945678 hasConceptScore W3129945678C121332964 @default.
- W3129945678 hasConceptScore W3129945678C122280245 @default.
- W3129945678 hasConceptScore W3129945678C12267149 @default.
- W3129945678 hasConceptScore W3129945678C134306372 @default.
- W3129945678 hasConceptScore W3129945678C138187205 @default.
- W3129945678 hasConceptScore W3129945678C147168706 @default.
- W3129945678 hasConceptScore W3129945678C151201525 @default.
- W3129945678 hasConceptScore W3129945678C153180895 @default.
- W3129945678 hasConceptScore W3129945678C154945302 @default.
- W3129945678 hasConceptScore W3129945678C158622935 @default.
- W3129945678 hasConceptScore W3129945678C177148314 @default.
- W3129945678 hasConceptScore W3129945678C199360897 @default.
- W3129945678 hasConceptScore W3129945678C2524010 @default.
- W3129945678 hasConceptScore W3129945678C26517878 @default.
- W3129945678 hasConceptScore W3129945678C33923547 @default.