Zobrazeno 1 - 10
of 14
pro vyhledávání: '"Rami Al-Rfou"'
Behavior prediction models have proliferated in recent years, especially in the popular real-world robotics application of autonomous driving, where representing the distribution over possible futures of moving agents is essential for safe and comfor
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::06e5a9d5c2bf71cf23639bce502b9171
http://arxiv.org/abs/2206.03970
http://arxiv.org/abs/2206.03970
Publikováno v:
ACL/IJCNLP (2)
Recently, mT5 - a massively multilingual version of T5 - leveraged a unified text-to-text format to attain state-of-the-art results on a wide variety of multilingual NLP tasks. In this paper, we investigate the impact of incorporating parallel data i
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::6ecc502f36493adff335cae7874f1a9d
http://arxiv.org/abs/2106.02171
http://arxiv.org/abs/2106.02171
Autor:
Noah Constant, Aditya Siddhant, Aditya Barua, Linting Xue, Mihir Kale, Adam Roberts, Rami Al-Rfou, Colin Raffel
Publikováno v:
NAACL-HLT
The recent "Text-to-Text Transfer Transformer" (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 th
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::6841f14537777a14674607884ca00048
https://doi.org/10.18653/v1/2021.naacl-main.41
https://doi.org/10.18653/v1/2021.naacl-main.41
There has been growing interest in parameter-efficient methods to apply pre-trained language models to downstream tasks. Building on the Prompt Tuning approach of Lester et al. (2021), which learns task-specific soft prompts to condition a frozen pre
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::51e25df605f7719510deba002e8ed53a
Knowledge Graph Based Synthetic Corpus Generation for Knowledge-Enhanced Language Model Pre-training
Publikováno v:
NAACL-HLT
Prior work on Data-To-Text Generation, the task of converting knowledge graph (KG) triples into natural text, focused on domain-specific benchmark datasets. In this paper, however, we verbalize the entire English Wikidata KG, and discuss the unique c
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::b5bb6865d8bb8ce3837661e8cf7df79c
http://arxiv.org/abs/2010.12688
http://arxiv.org/abs/2010.12688
Publikováno v:
EMNLP (1)
We present LAReQA, a challenging new benchmark for language-agnostic answer retrieval from a multilingual candidate pool. Unlike previous cross-lingual tasks, LAReQA tests for ``strong'' cross-lingual alignment, requiring semantically related \textit
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::432f6cbf60bea22e85af0a16aab03336
http://arxiv.org/abs/2004.05484
http://arxiv.org/abs/2004.05484
Publikováno v:
Analyses of Social Issues and Public Policy. 18:323-352
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::4227e50bba879eb689b07b8d69f6d3f4
https://doi.org/10.1111/asap.12158
https://doi.org/10.1111/asap.12158
Publikováno v:
The World Wide Web Conference.
Can neural networks learn to compare graphs without feature engineering? In this paper, we show that it is possible to learn representations for graph similarity with neither domain knowledge nor supervision (i.e.\ feature engineering or labeled grap
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::253046bc8aa8983f66ca0635d0188ef4
https://doi.org/10.1145/3308558.3313668
https://doi.org/10.1145/3308558.3313668
Publikováno v:
AAAI
LSTMs and other RNN variants have shown strong performance on character-level language modeling. These models are typically trained using truncated backpropagation through time, and it is common to assume that their success stems from their ability t
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::c0ebf60430a95dc087d9d118945992ba
http://arxiv.org/abs/1808.04444
http://arxiv.org/abs/1808.04444
Publikováno v:
CIKM
We propose a new method for embedding graphs while preserving directed edge information. Learning such continuous-space vector representations (or embeddings) of nodes in a graph is an important first step for using network information (from social n
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ac9d89a0e27bbf2f5eeb498ffb5f0f44
https://doi.org/10.1145/3132847.3132959
https://doi.org/10.1145/3132847.3132959