Search

In order to tackle the problem of abstractive summarization of long multi-sentence texts, it is critical to construct an efficient model, which can learn and represent multiple compositionalities better. In this paper, we introduce a temporal hierarchical pointer generator network that can represent multiple compositionalities in order to handle longer sequences of texts with a deep structure. We demonstrate how a multilayer gated recurrent neural network organizes itself with the help of an adaptive timescale in order to represent the compositions. The temporal hierarchical network is implemented with a multiple timescale architecture where the timescale of each layer is also learned during the training process through error backpropagation through time. We evaluate our proposed model using an Introduction-Abstract summarization dataset from scientific articles and the CNN/Daily Mail summarization benchmark dataset. The results illustrate that, we successfully implement a summary generation system for long texts by using the multiple timescale with adaptation concept. We also show that we have improved the summary generation system with our proposed model on the benchmark dataset.

Recently, intelligent dialog systems and smart assistants have attracted the attention of many, and development of novel dialogue agents have become a research challenge. Intelligent agents that can handle both domain-specific task-oriented and open-domain chit-chat dialogs are one of the major requirements in the current systems. In order to address this issue and to realize such smart hybrid dialogue systems, we develop a model to discriminate user utterance between task-oriented and chit-chat conversations. We introduce a hybrid of convolutional neural network (CNN) and a lateral multiple timescale gated recurrent units (LMTGRU) that can represent multiple temporal scale dependencies for the discrimination task. With the help of the combined slow and fast units of the LMTGRU, our model effectively determines whether a user will have a chit-chat conversation or a task-specific conversation with the system. We also show that the LMTGRU structure helps the model to perform well on longer text inputs. We address the lack of dataset by constructing a dataset using Twitter and Maluuba Frames data. The results of the experiments demonstrate that the proposed hybrid network outperforms the conventional models on the chat discrimination task as well as performed comparable to the baselines on various benchmark datasets.

Image transformation between multiple domains has become a challenging problem in deep generative networks. This is because, in real-world applications, finding paired images in different domains is an expensive and impractical task. This paper proposes a new model named joint moment-matching autoencoders(JMA). This model learns to perform cross-domain transformation over multiple domains based on perceptual loss and maximum mean discrepancy criteria, in the absence of any paired images between the domains. Our results show that the proposed JMA framework successfully learns to transform images between domains without any paired data. We demonstrate that our model has good performance in the generative context as well as in the domain transformation tasks with better computational efficiency than conventional methods.

In deep generative networks, one of the major challenges is to generate non-blurry, clearer images. Unlike the generative adversarial networks, generative models such as variational autoencoders, generative moment matching networks etc. use pixel-wise loss which leads to the generation of blurry images. In this paper, we propose an improved generative model called Generative Moment Matching Autoencoder (GMMA) with a feature-wise loss mechanism. We use a pre-trained VGGNet convolutional neural network to compute the loss at the various feature extraction layers. We evaluate the performance of our model on the MNIST and the Large-scale CelebFaces Attributes (CelebA) dataset. Our generative model outperforms the existing models on the log-likelihood estimation test. We also illustrate the effectiveness of our mechanism and the improved generation and reconstruction capabilities. The proposed GMMA with perceptual loss successfully alleviates the problem of blurry image generation.

A novel character-level neural language model is proposed in this paper. The proposed model incorporates a biologically inspired temporal hierarchy in the architecture for representing multiple compositions of language in order to handle longer sequences for the character-level language model. The temporal hierarchy is introduced in the language model by utilizing a Gated Recurrent Neural Network with multiple timescales. The proposed model incorporates a timescale adaptation mechanism for enhancing the performance of the language model. We evaluate our proposed model using the popular Penn Treebank and Text8 corpora. The experiments show that the use of multiple timescales in a Neural Language Model (NLM) enables improved performance despite having fewer parameters and with no additional computation requirements. Our experiments also demonstrate the ability of the adaptive temporal hierarchies to represent multiple compositonality without the help of complex hierarchical architectures and shows that better representation of the longer sequences lead to enhanced performance of the probabilistic language model.