Zobrazeno 1 - 10
of 37
pro vyhledávání: '"Hiroshi Noji"'
Publikováno v:
Journal of Natural Language Processing. 28:938-964
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::ee9ee611d208c5a6f65b727d6e58f247
https://doi.org/10.5715/jnlp.28.938
https://doi.org/10.5715/jnlp.28.938
Autor:
Hayate Iso, Hiroshi Noji, Eiji Aramaki, Naoaki Okazaki, Yusuke Miyao, Ichiro Kobayashi, Hiroya Takamura, Yui Uehara, Tatsuya Ishigaki
Publikováno v:
Journal of Natural Language Processing. 27:599-626
We propose a data-to-text generation model with two modules, one for tracking and the other for text generation. Our tracking module selects and keeps track of salient information and memorizes which record has been mentioned. Our generation module g
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::8d7852b59a175f23f2fc794d13c708e4
https://doi.org/10.5715/jnlp.27.599
https://doi.org/10.5715/jnlp.27.599
In computational linguistics, it has been shown that hierarchical structures make language models (LMs) more human-like. However, the previous literature has been agnostic about a parsing strategy of the hierarchical models. In this paper, we investi
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ce08d757b5b97b65598bb77a530cc108
http://arxiv.org/abs/2109.04939
http://arxiv.org/abs/2109.04939
Publikováno v:
AAAI
In logic-based approaches to reasoning tasks such as Recognizing Textual Entailment (RTE), it is important for a system to have a large amount of knowledge data. However, there is a tradeoff between adding more knowledge data for improved RTE perform
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::33a3913f1360c58365a1566ff968b323
https://doi.org/10.1609/aaai.v33i01.33017410
https://doi.org/10.1609/aaai.v33i01.33017410
Autor:
Yohei Oseki, Hiroshi Noji
Publikováno v:
ACL/IJCNLP (Findings)
As a language model that integrates traditional symbolic operations and flexible neural representations, recurrent neural network grammars (RNNGs) have attracted great attention from both scientific and engineering perspectives. However, RNNGs are kn
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::28e4a2b99d84e9692afa87d940741940
Autor:
Pierre Zweigenbaum, Hiroshi Noji, Olivier Ferret, Thomas Lavergne, Hicham El Boukkouri, Jun'ichi Tsujii
Publikováno v:
International Conference on Computational Linguistics
International Conference on Computational Linguistics, Dec 2020, Barcelona (on line), Spain. pp.6903-6915
HAL
COLING
International Conference on Computational Linguistics, Dec 2020, Barcelona (on line), Spain. pp.6903-6915
HAL
COLING
Due to the compelling improvements brought by BERT, many recent representation models adopted the Transformer architecture as their main building block, consequently inheriting the wordpiece tokenization system despite it not being intrinsically link
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::4cf5c14cc24c1e21ca88db1f9a2dc659
https://hal.archives-ouvertes.fr/hal-03100665/document
https://hal.archives-ouvertes.fr/hal-03100665/document
Publikováno v:
COLING
In this paper, we evaluate the progress of our field toward solving simple factoid questions over a knowledge base, a practically important problem in natural language interface to database. As in other natural language understanding tasks, a common
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::4076f00f591c80add9e350cd6ed84c48
https://doi.org/10.18653/v1/2020.coling-main.465
https://doi.org/10.18653/v1/2020.coling-main.465
Autor:
Kasumi Aoki, Yusuke Miyao, Ichiro Kobayashi, Hiroya Takamura, Keiichi Goshima, Tatsuya Ishigaki, Yui Uehara, Hiroshi Noji
Publikováno v:
COLING
Existing models for data-to-text tasks generate fluent but sometimes incorrect sentences e.g., “Nikkei gains” is generated when “Nikkei drops” is expected. We investigate models trained on contrastive examples i.e., incorrect sentences or ter
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::1172b4201899bab8e51b9751278c6ac8
https://doi.org/10.18653/v1/2020.coling-main.213
https://doi.org/10.18653/v1/2020.coling-main.213
Publikováno v:
ACL (1)
We propose a new domain adaptation method for Combinatory Categorial Grammar (CCG) parsing, based on the idea of automatic generation of CCG corpora exploiting cheaper resources of dependency trees. Our solution is conceptually simple, and not relyin
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::734f363fe3a39756e10b8f06db62cdbc
Autor:
Tatsuya Ishigaki, Eiji Aramaki, Hiroshi Noji, Yui Uehara, Yusuke Miyao, Naoaki Okazaki, Hiroya Takamura, Hayate Iso, Ichiro Kobayashi
Publikováno v:
Scopus-Elsevier
We propose a data-to-text generation model with two modules, one for tracking and the other for text generation. Our tracking module selects and keeps track of salient information and memorizes which record has been mentioned. Our generation module g
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::a06b61aa44a47237ab09c85f483ca35e