Matches in SemOpenAlex for { <https://semopenalex.org/work/W2952203969> ?p ?o ?g. }
- W2952203969 abstract "Recurrent neural network grammars (RNNG) are generative models of language which jointly model syntax and surface structure by incrementally generating a syntax tree and sentence in a top-down, left-to-right order. Supervised RNNGs achieve strong language modeling and parsing performance, but require an annotated corpus of parse trees. In this work, we experiment with unsupervised learning of RNNGs. Since directly marginalizing over the space of latent trees is intractable, we instead apply amortized variational inference. To maximize the evidence lower bound, we develop an inference network parameterized as a neural CRF constituency parser. On language modeling, unsupervised RNNGs perform as well their supervised counterparts on benchmarks in English and Chinese. On constituency grammar induction, they are competitive with recent neural language models that induce tree structures from words through attention mechanisms." @default.
- W2952203969 created "2019-06-27" @default.
- W2952203969 creator A5039107666 @default.
- W2952203969 creator A5045913119 @default.
- W2952203969 creator A5061131385 @default.
- W2952203969 creator A5067222034 @default.
- W2952203969 creator A5069646529 @default.
- W2952203969 creator A5085355324 @default.
- W2952203969 date "2019-04-07" @default.
- W2952203969 modified "2023-09-23" @default.
- W2952203969 title "Unsupervised Recurrent Neural Network Grammars" @default.
- W2952203969 cites W1557568992 @default.
- W2952203969 cites W1583697620 @default.
- W2952203969 cites W1632114991 @default.
- W2952203969 cites W179875071 @default.
- W2952203969 cites W1894835849 @default.
- W2952203969 cites W1959608418 @default.
- W2952203969 cites W2006036383 @default.
- W2952203969 cites W2020382207 @default.
- W2952203969 cites W2047706513 @default.
- W2952203969 cites W2076083849 @default.
- W2952203969 cites W2083195487 @default.
- W2952203969 cites W2096204319 @default.
- W2952203969 cites W2103149536 @default.
- W2952203969 cites W2109926133 @default.
- W2952203969 cites W2111369166 @default.
- W2952203969 cites W2115803233 @default.
- W2952203969 cites W2122262818 @default.
- W2952203969 cites W2129882630 @default.
- W2952203969 cites W2134036914 @default.
- W2952203969 cites W2134275125 @default.
- W2952203969 cites W2138433202 @default.
- W2952203969 cites W2152561660 @default.
- W2952203969 cites W2154368244 @default.
- W2952203969 cites W2156024017 @default.
- W2952203969 cites W2157762871 @default.
- W2952203969 cites W2162016438 @default.
- W2952203969 cites W2164151151 @default.
- W2952203969 cites W2250520759 @default.
- W2952203969 cites W2250861254 @default.
- W2952203969 cites W2251299535 @default.
- W2952203969 cites W2464234964 @default.
- W2952203969 cites W2512597464 @default.
- W2952203969 cites W2554915555 @default.
- W2952203969 cites W2586050494 @default.
- W2952203969 cites W2594047108 @default.
- W2952203969 cites W2619269479 @default.
- W2952203969 cites W2744831323 @default.
- W2952203969 cites W2791751435 @default.
- W2952203969 cites W2798727047 @default.
- W2952203969 cites W2804228899 @default.
- W2952203969 cites W2859444450 @default.
- W2952203969 cites W2883510303 @default.
- W2952203969 cites W2888922637 @default.
- W2952203969 cites W2889009749 @default.
- W2952203969 cites W2889260178 @default.
- W2952203969 cites W2889404673 @default.
- W2952203969 cites W2889830342 @default.
- W2952203969 cites W2890627034 @default.
- W2952203969 cites W2890908793 @default.
- W2952203969 cites W2897767292 @default.
- W2952203969 cites W2901751252 @default.
- W2952203969 cites W2918073309 @default.
- W2952203969 cites W2932376173 @default.
- W2952203969 cites W2950919476 @default.
- W2952203969 cites W2951714314 @default.
- W2952203969 cites W2953130735 @default.
- W2952203969 cites W2962695963 @default.
- W2952203969 cites W2962728167 @default.
- W2952203969 cites W2962788148 @default.
- W2952203969 cites W2962832505 @default.
- W2952203969 cites W2962897886 @default.
- W2952203969 cites W2962934478 @default.
- W2952203969 cites W2962941914 @default.
- W2952203969 cites W2962946519 @default.
- W2952203969 cites W2963073938 @default.
- W2952203969 cites W2963084773 @default.
- W2952203969 cites W2963135265 @default.
- W2952203969 cites W2963223306 @default.
- W2952203969 cites W2963319176 @default.
- W2952203969 cites W2963347649 @default.
- W2952203969 cites W2963351648 @default.
- W2952203969 cites W2963411763 @default.
- W2952203969 cites W2963451457 @default.
- W2952203969 cites W2963471154 @default.
- W2952203969 cites W2963525616 @default.
- W2952203969 cites W2963537482 @default.
- W2952203969 cites W2963580443 @default.
- W2952203969 cites W2963617989 @default.
- W2952203969 cites W2963735467 @default.
- W2952203969 cites W2963748792 @default.
- W2952203969 cites W2963754491 @default.
- W2952203969 cites W2963808856 @default.
- W2952203969 cites W2963863610 @default.
- W2952203969 cites W2963899396 @default.
- W2952203969 cites W2964030814 @default.
- W2952203969 cites W2964121744 @default.
- W2952203969 cites W2964199361 @default.
- W2952203969 cites W2964232608 @default.
- W2952203969 cites W2964263959 @default.