-
1“...Multi-task learning (MTL) in deep neural networks for NLP has recently received increasing interest due to some compelling benefits, including its potential to...”
-
2“...Machine Learning seeks to identify and encode bodies of knowledge within provided datasets. However, data encodes subjective content, which determines the...”
-
3“...Provider: Deutsche Digitale Bibliothek - Institution: Institut für deutsche Sprache - Bibliothek - Data provided by Europeana Collections- The task-oriented...”
-
4“...Provider: Deutsche Digitale Bibliothek - Institution: Institut für deutsche Sprache - Bibliothek - Data provided by Europeana Collections- We describe a...”
-
5by Ruder, Sebastian Bingel, Joachim Augenstein, Isabelle Søgaard, Anders Published in Proceedings of the ... AAAI Conference on Artificial Intelligence (17.07.2019)“...Multi-task learning (MTL) allows deep neural networks to learn from related tasks by sharing parameters with other networks. In practice, however, MTL involves...”
-
6“...Multi-task learning (MTL) in deep neural networks for NLP has recently received increasing interest due to some compelling benefits, including its potential to...”
-
7by Bingel, Joachim Hansen, Victor Petrén Bach Gonzalez, Ana Valeria Budzianowski, Paweł Augenstein, Isabelle Søgaard, Anders“...Task oriented dialogue systems rely heavily on specialized dialogue state tracking (DST) modules for dynamically predicting user intent throughout the...”
16.09.2019
-
8“...Multi-task learning (MTL) allows deep neural networks to learn from related tasks by sharing parameters with other networks. In practice, however, MTL involves...”
-
9“...Provider: Deutsche Digitale Bibliothek - Institution: Institut für deutsche Sprache - Bibliothek - Data provided by Europeana Collections- KoralQuery is a...”
-
10“...Accurately detecting hate speech using supervised classification is dependent on data that is annotated by humans. Attaining high agreement amongst annotators...”
-
11by Bingel, Joachim“...Provider: Deutsche Digitale Bibliothek - Institution: Institut für deutsche Sprache - Bibliothek - Data provided by Europeana Collections- The present thesis...”
16.10.2015
-
12“...Learning attention functions requires large volumes of data, but many NLP tasks simulate human behavior, and in this paper, we show that human attention really...”
-
13by Bański, Piotr Bingel, Joachim Diewald, Nils Hanl, Michael Kupietz, Marc Margaretha, Eliza Witt, Andreas“...Provider: Deutsche Digitale Bibliothek - Institution: Institut für deutsche Sprache - Bibliothek - Data provided by Europeana Collections- KorAP is a corpus...”
01.07.2016
-
14by Bingel, Joachim Hansen, Victor Petrén Bach Gonzalez, Ana Valeria Budzianowski, Paweł Augenstein, Isabelle Søgaard, Anders“...Task oriented dialogue systems rely heavily on specialized dialogue state tracking (DST) modules for dynamically predicting user intent throughout the...”
16.09.2019
-
15“...Multi-task learning (MTL) allows deep neural networks to learn from related tasks by sharing parameters with other networks. In practice, however, MTL involves...”
-
16“...This work describes the system presented by the CoAStaL Natural Language Processing group at University of Copenhagen. The main system we present uses the same...”
-
17“...For many of the world’s languages, there are no or very few linguistically annotated resources. On the other hand, raw text, and often also dictionaries, can...”
-
18“...Learning attention functions requires largevolumes of data, but many NLP tasks simulatehuman behavior, and in this paper, weshow that human attention really...”
-
19by Bański, Piotr Bingel, Joachim Diewald, Nils Frick, Elena Hanl, Michael Kupietz, Marc Pȩzik, Piotr Schnober, Carsten Witt, Andreas“...Provider: Deutsche Digitale Bibliothek - Institution: Institut für deutsche Sprache - Bibliothek - Data provided by Europeana Collections- The KorAP project...”
18.11.2014
-
20by Alva-Manchego, Fernando Bingel, Joachim Paetzold, Gustavo Henrique Scarton, Carolina Specia, Lucia“...Current research in text simplification has been hampered by two central problems: (i) the small amount of high-quality parallel simplification data available,...”
27.11.2017