Zobrazeno 1 - 10
of 20
pro vyhledávání: '"Dani Yogatama"'
Autor:
Devendra Singh Sachan, Mike Lewis, Dani Yogatama, Luke Zettlemoyer, Joelle Pineau, Manzil Zaheer
Publikováno v:
Transactions of the Association for Computational Linguistics, Vol 11, Pp 600-616 (2023)
AbstractWe introduce ART, a new corpus-level autoencoding approach for training dense retrieval models that does not require any labeled training data. Dense retrieval is a central challenge for open-domain tasks, such as Open QA, where state-of-the-
Externí odkaz:
https://doaj.org/article/1876b586e8df46f39c701449c9e93aa5
Publikováno v:
Transactions of the Association for Computational Linguistics, Vol 9, Pp 362-373 (2021)
AbstractWe present a language model that combines a large parametric neural network (i.e., a transformer) with a non-parametric episodic memory component in an integrated architecture. Our model uses extended short-term context by caching local hidde
Externí odkaz:
https://doaj.org/article/4d6a91c6e8ff40bfb1af0dd1fd3c888a
Publikováno v:
Transactions of the Association for Computational Linguistics. 10:555-572
We present a memory-augmented approach to condition an autoregressive language model on a knowledge graph. We represent the graph as a collection of relation triples and retrieve relevant relations for a given context to improve text generation. Expe
Publikováno v:
Proceedings of the International AAAI Conference on Web and Social Media. 7:737-740
Who influences a politician’s public statements? In this paper, we explore one plausible explanation: that financial incentives from campaign contributors affect what politicians say. Based on this idea, we design text-driven models for campaign co
Autor:
Daniel Fried, Phil Blunsom, Lingpeng Kong, Chris Dyer, Adhiguna Kuncoro, Laura Rimell, Dani Yogatama
Publikováno v:
Transactions of the Association for Computational Linguistics. 8:776-794
Textual representation learners trained on large amounts of data have achieved notable success on downstream tasks; intriguingly, they have also performed well on challenging tests of syntactic competence. Given this success, it remains an open quest
Autor:
Tom Schaul, David Silver, James Molloy, Junhyuk Oh, Katrina McKinney, Oriol Vinyals, David H. Choi, Junyoung Chung, Tobias Pohlen, Dani Yogatama, Tobias Pfaff, Demis Hassabis, Michael Mathieu, Dan Horgan, Ivo Danihelka, Igor Babuschkin, Dario Wünsch, Tom Le Paine, Yury Sulsky, Wojciech Marian Czarnecki, Rémi Leblond, Ziyu Wang, Andrew Dudzik, Trevor Cai, Chris Apps, Yuhuai Wu, David Budden, Valentin Dalibard, Timo Ewalds, Oliver Smith, John P. Agapiou, Aja Huang, Roman Ring, Petko Georgiev, Max Jaderberg, Koray Kavukcuoglu, Alexander Vezhnevets, Caglar Gulcehre, Manuel Kroiss, Laurent Sifre, Richard E. Powell, Timothy P. Lillicrap
Publikováno v:
Nature. 575:350-354
Many real-world applications require artificial agents to compete and coordinate with other agents in complex environments. As a stepping stone to this goal, the domain of StarCraft has emerged as an important challenge for artificial intelligence re
Autor:
Jungo Kasai, Hao Peng, Yizhe Zhang, Dani Yogatama, Gabriel Ilharco, Nikolaos Pappas, Yi Mao, Weizhu Chen, Noah A. Smith
Transformers have outperformed recurrent neural networks (RNNs) in natural language generation. But this comes with a significant computational cost, as the attention mechanism's complexity scales quadratically with sequence length. Efficient transfo
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::a491a5ce8fd5f97c8088051a5d5a8eb2
http://arxiv.org/abs/2103.13076
http://arxiv.org/abs/2103.13076
Publikováno v:
ACL
We review motivations, definition, approaches, and methodology for unsupervised cross-lingual learning and call for a more rigorous position in each of them. An existing rationale for such research is based on the lack of parallel data for many of th
Autor:
Pushmeet Kohli, Ray Jiang, Jack W. Rae, Dani Yogatama, Robert Stanforth, Po-Sen Huang, Huan Zhang, Vishal Maini, Johannes Welbl
Publikováno v:
EMNLP (Findings)
Advances in language modeling architectures and the availability of large text corpora have driven progress in automatic text generation. While this results in models capable of generating coherent texts, it also prompts models to internalize social
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::6b530aab75c46c96120f2006913ca2bd
http://arxiv.org/abs/1911.03064
http://arxiv.org/abs/1911.03064
Autor:
Pushmeet Kohli, Krishnamurthy Dvijotham, Chris Dyer, Sven Gowal, Dani Yogatama, Johannes Welbl, Robert Stanforth, Po-Sen Huang
Publikováno v:
EMNLP/IJCNLP (1)
Neural networks are part of many contemporary NLP systems, yet their empirical successes come at the price of vulnerability to adversarial attacks. Previous work has used adversarial training and data augmentation to partially mitigate such brittlene