“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

Automatic speech recognition, especially large vocabulary continuous speech recognition, is an important issue in the field of machine learning. For a long time, the hidden Markov model (HMM)-Gaussian mixed model (GMM) has been the mainstream speech recognition framework. But recently, HMM-deep neural network (DNN) model and the end-to-end model using deep learning has achieved performance beyond HMM-GMM. Both using deep learning techniques, these two models have comparable performances. However, the HMM-DNN model itself is limited by various unfavorable factors such as data forced segmentation alignment, independent hypothesis, and multi-module individual training inherited from HMM, while the end-to-end model has a simplified model, joint training, direct output, no need to force data alignment and other advantages. Therefore, the end-to-end model is an important research direction of speech recognition. In this paper we review the development of end-to-end model. This paper first introduces the basic ideas, advantages and disadvantages of HMM-based model and end-to-end models, and points out that end-to-end model is the development direction of speech recognition. Then the article focuses on the principles, progress and research hotspots of three different end-to-end models, which are connectionist temporal classification (CTC)-based, recurrent neural network (RNN)-transducer and attention-based, and makes theoretically and experimentally detailed comparisons. Their respective advantages and disadvantages and the possible future development of the end-to-end model are finally pointed out. Automatic speech recognition is a pattern recognition task in the field of computer science, which is a subject area of Symmetry.

An Overview of End-to-End Automatic Speech Recognition

Deep neural networks have attained great success in handling high dimensional data, especially images. However, generating naturalistic images containing ginormous subjects for different tasks like image classification, segmentation, object detection, reconstruction, etc., is continued to be a difficult task. Generative modelling has the potential to learn any kind of data distribution in an unsupervised manner. Variational autoencoder (VAE), autoregressive models, and generative adversarial network (GAN) are the popular generative modelling approaches that generate data distributions. Among these, GANs have gained much attention from the research community in recent years in terms of generating quality images and data augmentation. In this context, we collected research articles that employed GANs for solving various tasks from popular databases and summarized them based on their application. The main objective of this article is to present the nuts and bolts of GANs, state-of-the-art related work and its applications, evaluation metrics, challenges involved in training GANs, and benchmark datasets that would benefit naive and enthusiastic researchers who are interested in working on GANs.

Generative adversarial networks: a survey on applications and challenges

During a new disease outbreak, frustration and uncertainties among affected and vulnerable population increase. Affected communities look for known symptoms, prevention measures, and treatment strategies. On the other hand, health organizations try to get situational updates to assess the severity of the outbreak, known affected cases, and other details. Recent emergence of social media platforms such as Twitter provide convenient ways and fast access to disseminate and consume information to/from a wider audience. Research studies have shown potential of this online information to address information needs of concerned authorities during outbreaks, epidemics, and pandemics. In this work, we target three types of end-users (i) vulnerable population-people who are not yet affected and are looking for prevention related information (ii) affected population-people who are affected and looking for treatment related information, and (iii) health organizations-like WHO, who are interested in gaining situational awareness to make timely decisions. We use Twitter data from two recent outbreaks (Ebola and MERS) to build an automatic classification approach useful to categorize tweets into different disease related categories. Moreover, the classified messages are used to generate different kinds of summaries useful for affected and vulnerable communities as well as health organizations. Results obtained from extensive experimentation show the effectiveness of the proposed approach.

/pdf/classifying-and-summarizing-information-from-microblogs-4plo1hv82k.pdf

Classifying and Summarizing Information from Microblogs During Epidemics

At the outbreak of an epidemic, affected communities want/need to get aware of disease symptoms, preventive measures, and treatment strategies. On the other hand, health organizations try to get situational updates to assess the severity of the outbreak, known affected cases, and other details. Recent emergence of social media platforms such as Twitter provide convenient ways and fast access to disseminate and consume information to/from a wider audience. Research studies have shown potential of this online information to address information needs of concerned authorities during outbreaks, epidemics, and pandemics. In this work, we target three communities (i) people who are not affected yet and are looking for prevention-related information (ii) people who are affected and looking for treatment-related information, and (iii) health organizations like WHO, who are interested in gaining situational awareness to make timely decisions. We use Twitter data from two recent outbreaks (Ebola and MERS) to built an automatic classification approach using low level lexical features which are useful to categorize tweets into different disease-related categories.

Classifying Information from Microblogs during Epidemics

Community question answering (CQA) services have emerged as a type of popular social platforms. In the social network, experts provide knowledgeable answers to the questions in their domain of expertise, while celebrities publish influential opinions toward the topics led by some questions. Given the large amount of knowledge organized in the format of question–answers, an interesting research problem is to recommend questions to users so as to maximize their engagements with the platform. However, recommending questions in CQA services is a non-trivial task. Data sources in the CQA services are of different types. It is challenging to incorporate heterogeneous information for the recommendation task. Furthermore, data sparsity is an inherent problem in such platforms. In this paper, we propose a model that is able to jointly model both implicit and explicit information for question recommendation. The model integrates multiple data sources and addresses the problem of data heterogeneity. In the proposed model, we dynamically discover latent user groups and incorporate those hierarchical information to bridge the semantic gaps among users in the shared latent space. We evaluate the proposed model on two real-world datasets, and demonstrate that our model outperforms the state-of-the-art alternatives by a large margin. We also investigate different structures of the proposed model to study the effects of different data sources.

Joint Implicit and Explicit Neural Networks for Question Recommendation in CQA Services

Recently, text style transfer has become a very hot research topic in the field of natural language processing. However, the conventional text style transfer is unidirectional, and it is not possible to obtain a model with multidirectional transformations through training once. To address this limitation, we propose a new task called multidirectional text style transfer. It aims to use a single model to transfer the underlying style of text among multiple style attributes and keep its main content unchanged. In this paper, we propose Unified Generative Adversarial Networks (UGAN), a practical approach that combines target vector and generative adversarial techniques to perform multidirectional text style transfer. Our model allows simultaneous training of multi-attribute data on a single network. Such unified structure makes our model more efficient and flexible than existing approaches. We demonstrate the superiority of our approach on three benchmark datasets. Experimental results show that our method not only outperforms other baselines, but also reduces training time by an average of 13%.

/pdf/ugan-unified-generative-adversarial-networks-for-5qrt5plkdn.pdf

UGAN: Unified Generative Adversarial Networks for Multidirectional Text Style Transfer

Temporal relation extraction aims to extract temporal relations between event pairs, which is crucial for natural language understanding. Few efforts have been devoted to capturing the global features. In this paper, we propose RSGT: Relational Structure Guided Temporal Relation Extraction to extract the relational structure features that can fit for both inter-sentence and intra-sentence relations. Specifically, we construct a syntactic-and-semantic-based graph to extract relational structures. Then we present a graph neural network based model to learn the representation of this graph. After that, an auxiliary temporal neighbor prediction task is used to fine-tune the encoder to get more comprehensive node representations. Finally, we apply a conflict detection and correction algorithm to adjust the wrongly predicted labels. Experiments on two well-known datasets, MATRES and TB-Dense, demonstrate the superiority of our method (2.3% F1 improvement on MATRES, 3.5% F1 improvement on TB-Dense).

RSGT: Relational Structure Guided Temporal Relation Extraction

Health forums have gained attention from researchers for studying various topics on healthcare. In many of these studies, identifying biomedical words by using the MetaMap is often a pre-processing step. MetaMap is a popular tool for recognizing Unified Medical Language System (UMLS) concepts in free text. However, MetaMap favors identifying terminologies used by professionals rather than laymen terms by the common users. The word labels given by MetaMap on social media may not be accurate, and may adversely affect the next level studies. In this study, we manually annotate the correctness of medical words extracted by MetaMap from 100 posts in HealthBoards and get a precision of 43.75 %. We argue that directly applying MetaMap on social media data in healthcare may not be a good choice for identifying the medical words.

When MetaMap Meets Social Media in Healthcare: Are the Word Labels Correct?

Based on the natural truth that similar questions correspond to similar SQL queries, a CBR-based approach is proposed to deal with the Text-to-SQL task in this paper. We follow the traditional CBR processes: similarity assessment, case retrieval, and case reuse. First, we introduce a neural classifier in the similarity assessment stage and comprehensively uses classification probability and literal cosine similarity to measure similarity. Then, based on the results of the similarity assessment, our model retrieves a case template. Finally, our model fills the columns and values generated by the Ranker module and Question Answering (QA) module into the solution template. At this point, a SQL query suitable for the new case is generated. We evaluate our models on a large-scale Text-to-SQL dataset—WikiSQL. Experimentally, our model has a competitive performance compared with the baseline and significantly improves the accuracy of the aggregation function prediction.

https://link.springer.com/content/pdf/10.1007%2F978-3-030-86957-1_20.pdf

Xiaodong Wang

Papers

Joint Implicit and Explicit Neural Networks for Question Recommendation in CQA Services

UGAN: Unified Generative Adversarial Networks for Multidirectional Text Style Transfer

RSGT: Relational Structure Guided Temporal Relation Extraction

When MetaMap Meets Social Media in Healthcare: Are the Word Labels Correct?

Similar Questions Correspond to Similar SQL Queries: A Case-Based Reasoning Approach for Text-to-SQL Translation