Matches in SemOpenAlex for { <https://semopenalex.org/work/W3157156950> ?p ?o ?g. }
- W3157156950 abstract "Abstract Probabilistic language models, e.g. those based on recurrent neural networks such as long short-term memory models (LSTMs), often face the problem of finding a high probability prediction from a sequence of random variables over a set of tokens. This is commonly addressed using a form of greedy decoding such as beam search, where a limited number of highest-likelihood paths (the beam width) of the decoder are kept, and at the end the maximum-likelihood path is chosen. In this work, we construct a quantum algorithm to find the globally optimal parse (i.e. for infinite beam width) with high constant success probability. When the input to the decoder follows a power law with exponent k > 0, our algorithm has runtime R n f ( R , k ) , where R is the alphabet size, n the input length; here f < 1/2, and $frightarrow 0$ <mml:math xmlns:mml=http://www.w3.org/1998/Math/MathML><mml:mi>f</mml:mi><mml:mo>→</mml:mo><mml:mn>0</mml:mn></mml:math> exponentially fast with increasing k , hence making our algorithm always more than quadratically faster than its classical counterpart. We further modify our procedure to recover a finite beam width variant, which enables an even stronger empirical speedup while still retaining higher accuracy than possible classically. Finally, we apply this quantum beam search decoder to Mozilla’s implementation of Baidu’s DeepSpeech neural net, which we show to exhibit such a power law word rank frequency." @default.
- W3157156950 created "2021-05-10" @default.
- W3157156950 creator A5000073123 @default.
- W3157156950 creator A5047557043 @default.
- W3157156950 creator A5074452192 @default.
- W3157156950 date "2021-04-30" @default.
- W3157156950 modified "2023-10-17" @default.
- W3157156950 title "A quantum search decoder for natural language processing" @default.
- W3157156950 cites W100226316 @default.
- W3157156950 cites W1492999010 @default.
- W3157156950 cites W15592790 @default.
- W3157156950 cites W1570445666 @default.
- W3157156950 cites W1600928351 @default.
- W3157156950 cites W1645247980 @default.
- W3157156950 cites W1860144983 @default.
- W3157156950 cites W1970961429 @default.
- W3157156950 cites W1972917397 @default.
- W3157156950 cites W1983360325 @default.
- W3157156950 cites W1996684294 @default.
- W3157156950 cites W2001959333 @default.
- W3157156950 cites W2001960495 @default.
- W3157156950 cites W2040792108 @default.
- W3157156950 cites W2048176942 @default.
- W3157156950 cites W2054801208 @default.
- W3157156950 cites W2067381074 @default.
- W3157156950 cites W2072607050 @default.
- W3157156950 cites W2076063813 @default.
- W3157156950 cites W2092477611 @default.
- W3157156950 cites W2118172634 @default.
- W3157156950 cites W2124479173 @default.
- W3157156950 cites W2134036914 @default.
- W3157156950 cites W2137147061 @default.
- W3157156950 cites W2140505514 @default.
- W3157156950 cites W2141058327 @default.
- W3157156950 cites W2174510006 @default.
- W3157156950 cites W2257937122 @default.
- W3157156950 cites W2296217176 @default.
- W3157156950 cites W2487259721 @default.
- W3157156950 cites W2511719953 @default.
- W3157156950 cites W2567469948 @default.
- W3157156950 cites W2759506558 @default.
- W3157156950 cites W2766294686 @default.
- W3157156950 cites W2897739530 @default.
- W3157156950 cites W2905315630 @default.
- W3157156950 cites W2908019568 @default.
- W3157156950 cites W2923481626 @default.
- W3157156950 cites W2963088785 @default.
- W3157156950 cites W2963096510 @default.
- W3157156950 cites W2963341399 @default.
- W3157156950 cites W2963382396 @default.
- W3157156950 cites W2963620441 @default.
- W3157156950 cites W2963665552 @default.
- W3157156950 cites W2963951265 @default.
- W3157156950 cites W2964110616 @default.
- W3157156950 cites W2990961515 @default.
- W3157156950 cites W2995969307 @default.
- W3157156950 cites W3011600946 @default.
- W3157156950 cites W3096104971 @default.
- W3157156950 cites W3098280028 @default.
- W3157156950 cites W3100071985 @default.
- W3157156950 cites W3100604492 @default.
- W3157156950 cites W3101135395 @default.
- W3157156950 cites W3101299045 @default.
- W3157156950 cites W3102774135 @default.
- W3157156950 cites W4249179470 @default.
- W3157156950 cites W4253103663 @default.
- W3157156950 cites W4376522561 @default.
- W3157156950 cites W2122472280 @default.
- W3157156950 doi "https://doi.org/10.1007/s42484-021-00041-1" @default.
- W3157156950 hasPublicationYear "2021" @default.
- W3157156950 type Work @default.
- W3157156950 sameAs 3157156950 @default.
- W3157156950 citedByCount "5" @default.
- W3157156950 countsByYear W31571569502020 @default.
- W3157156950 countsByYear W31571569502021 @default.
- W3157156950 countsByYear W31571569502023 @default.
- W3157156950 crossrefType "journal-article" @default.
- W3157156950 hasAuthorship W3157156950A5000073123 @default.
- W3157156950 hasAuthorship W3157156950A5047557043 @default.
- W3157156950 hasAuthorship W3157156950A5074452192 @default.
- W3157156950 hasBestOaLocation W31571569501 @default.
- W3157156950 hasConcept C111919701 @default.
- W3157156950 hasConcept C11413529 @default.
- W3157156950 hasConcept C125583679 @default.
- W3157156950 hasConcept C137293760 @default.
- W3157156950 hasConcept C154945302 @default.
- W3157156950 hasConcept C19889080 @default.
- W3157156950 hasConcept C41008148 @default.
- W3157156950 hasConcept C57273362 @default.
- W3157156950 hasConcept C68339613 @default.
- W3157156950 hasConceptScore W3157156950C111919701 @default.
- W3157156950 hasConceptScore W3157156950C11413529 @default.
- W3157156950 hasConceptScore W3157156950C125583679 @default.
- W3157156950 hasConceptScore W3157156950C137293760 @default.
- W3157156950 hasConceptScore W3157156950C154945302 @default.
- W3157156950 hasConceptScore W3157156950C19889080 @default.
- W3157156950 hasConceptScore W3157156950C41008148 @default.
- W3157156950 hasConceptScore W3157156950C57273362 @default.
- W3157156950 hasConceptScore W3157156950C68339613 @default.
- W3157156950 hasFunder F4320321847 @default.