Results 21 to 30 of about 11,864 (109)

Revisiting Activation Regularization for Language RNNs [PDF]

arXiv, 2017
Recurrent neural networks (RNNs) serve as a fundamental building block for many sequence tasks across natural language processing. Recent research has focused on recurrent dropout techniques or custom RNN cells in order to improve performance. Both of these can require substantial modifications to the machine learning model or to the underlying RNN ...
arxiv  

Borel Vizing's Theorem for 2-Ended Groups [PDF]

arXiv, 2021
We show that Vizing's Theorem holds in the Borel context for graphs induced by actions of 2-ended groups, and ask whether it holds more generally for everywhere two ended Borel graphs.
arxiv  

Regularizing and Optimizing LSTM Language Models [PDF]

arXiv, 2017
Recurrent neural networks (RNNs), such as long short-term memory networks (LSTMs), serve as a fundamental building block for many sequence learning tasks, including machine translation, language modeling, and question answering. In this paper, we consider the specific problem of word-level language modeling and investigate strategies for regularizing ...
arxiv  

A Flexible Approach to Automated RNN Architecture Generation [PDF]

arXiv, 2017
The process of designing neural architectures requires expert knowledge and extensive trial and error. While automated architecture search may simplify these requirements, the recurrent neural network (RNN) architectures generated by existing methods are limited in both flexibility and components.
arxiv  

School evasion: A hard reality [PDF]

arXiv, 2008
The present work has as objective to show the profile of students who abandoned the studies in a High School, located in Sao Joao de Meriti city, municipal district of Rio de Janeiro state, by means of statistical analysis. The presented indices portray an undesirable reality with almost 20% school evasion, beyond showing that more the half of the ...
arxiv  

L'himnari SMV 62 de la seu de Mallorca. Descripció i proposta musical [PDF]

, 2017
Després d'una nova revisió de l'original de l'himnari SMV 62 de la seu de Mallorca, basant-nos en les obres dels 145 folis actuals, hem actualitzat i completat l'índex del seu contingut, complementat amb informació específica sobre els set tipus de ...
Romà Escalas i Llimona
core   +1 more source

Single Headed Attention RNN: Stop Thinking With Your Head [PDF]

arXiv, 2019
The leading approaches in language modeling are all obsessed with TV shows of my youth - namely Transformers and Sesame Street. Transformers this, Transformers that, and over here a bonfire worth of GPU-TPU-neuromorphic wafer scale silicon. We opt for the lazy path of old and proven techniques with a fancy crypto inspired acronym: the Single Headed ...
arxiv  

Dynamic Memory Networks for Visual and Textual Question Answering [PDF]

arXiv, 2016
Neural network architectures with memory and attention mechanisms exhibit certain reasoning capabilities required for question answering. One such architecture, the dynamic memory network (DMN), obtained high accuracy on a variety of language tasks. However, it was not shown whether the architecture achieves strong results for question answering when ...
arxiv  

Pointer Sentinel Mixture Models [PDF]

arXiv, 2016
Recent neural network sequence models with softmax classifiers have achieved their best language modeling performance only with very large hidden states and large vocabularies. Even then they struggle to predict rare or unseen words even if the context makes the prediction unambiguous.
arxiv  

Quasi-Recurrent Neural Networks [PDF]

arXiv, 2016
Recurrent neural networks are a powerful tool for modeling sequential data, but the dependence of each timestep's computation on the previous timestep's output limits parallelism and makes RNNs unwieldy for very long sequences. We introduce quasi-recurrent neural networks (QRNNs), an approach to neural sequence modeling that alternates convolutional ...
arxiv