检索结果-内蒙古大学图书馆

variational model for low-resource natural language generation in spoken dialogue systems

COMPUTER SPEECH AND LANGUAGE 2021年 65卷 101120-101120页

作者： Van-Khanh Tran Le-Minh Nguyen Japan Adv Inst Sci & Technol JAIST 1-1 Asahidai Nomi Ishikawa 9231292 Japan ICTU Thai Nguyen Univ Univ Informat & Commun Technol Thai Nguyen Vietnam

Natural Language Generation (NLG) plays a critical role in Spoken Dialogue Systems (SDSs), aims at converting a meaning representation into natural language utterances. Recent deep learning-based generators have shown improving results irrespective of providing sufficient annotated data. Nevertheless, how to build a generator that can effectively utilize as much of knowledge from a low-resource setting data is a crucial issue for NLG in SDSs. This paper presents a variational-based NLG framework to tackle the NLG problem of having limited annotated data in two scenarios, domain adaptation and low-resource in-domain training data. Based on this framework, we propose a novel adversarial domain adaptation NLG taclking the former issue, while the latter issue is also handled by a second proposed dual variational model. We extensively conducted the experiments on four different domains in a variety of training scenarios, in which the experimental results show that the proposed methods not only outperform previous methods when having sufficient training dataset but also show its ability to work acceptably well when there is a small amount of in-domain data or adapt quickly to a new domain with only a low-resource target domain data. (C) 2020 Published by Elsevier Ltd.

关键词： Neural language generation Domain adaptation Low-resource data variational autoencoder Deconvolutional neural network CNN RNN LSTM

来源：评论

学校读者我要写书评

暂无评论

GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies

引用

IEEE ACCESS 2021年 9卷 148994-149042页

作者： Ditzel, Carsten Dietmayer, Klaus Inst Measurement Control & Microtechnol D-89081 Ulm Germany

Autonomous systems require a continuous and dependable environment perception for navigation and decision-making, which is best achieved by combining different sensor types. Radar continues to function robustly in compromised circumstances in which cameras become impaired, guaranteeing a steady inflow of information. Yet, camera images provide a more intuitive and readily applicable impression of the world. This work combines the complementary strengths of both sensor types in a unique self-learning fusion approach for a probabilistic scene reconstruction in adverse surrounding conditions. After reducing the memory requirements of both high-dimensional measurements through a decoupled stochastic self-supervised compression technique, the proposed algorithm exploits similarities and establishes correspondences between both domains at different feature levels during training. Then, at inference time, relying exclusively on radio frequencies, the model successively predicts camera constituents in an autoregressive and self-contained process. These discrete tokens are finally transformed back into an instructive view of the respective surrounding, allowing to visually perceive potential dangers for important tasks downstream.

关键词： Radar Cameras Radar imaging Robot sensing systems Probabilistic logic Task analysis Signal processing algorithms Radar signal processing computer vision sensor fusion deep learning machine learning variational autoencoder self-supervised learning annotation-free learning scene understanding

来源：评论

学校读者我要写书评

暂无评论

Data-Driven Discovery of 2D Materials for Solar Water Splitting

引用

FRONTIERS IN MATERIALS 2021年 8卷

作者： Agarwal, Abhishek Goverapet Srinivasan, Sriram Rai, Beena Tata Consulatncy Serv Ltd TCS Res Pune Maharashtra India

Hydrogen economy, wherein hydrogen is used as the fuel in the transport and energy sectors, holds significant promise in mitigating the deleterious effects of global warming. Photocatalytic water splitting using sunlight is perhaps the cleanest way of producing the hydrogen fuel. Among various other factors, widespread adoption of this technology has mainly been stymied by the lack of a catalyst material with high efficiency. 2D materials have shown significant promise as efficient photocatalysts for water splitting. The availability of open databases containing the "computed" properties of 2D materials and advancements in deep learning now enable us to do "inverse" design of these 2D photocatalysts for water splitting. We use one such database (Jain et al., ACS Energ. Lett. 2019, 4, 6, 1410-1411) to build a generative model for the discovery of novel 2D photocatalysts. The structures of the materials were converted into a 3D image-based representation that was used to train a cell, a basis autoencoder and a segmentation network to ascertain the lattice parameters as well as position of atoms from the images. Subsequently, the cell and basis encodings were used to train a conditional variational autoencoder (CVAE) to learn a continuous representation of the materials in a latent space. The latent space of the CVAE was then sampled to generate several new 2D materials that were likely to be efficient photocatalysts for water splitting. The bandgap of the generated materials was predicted using a graph neural network model while the band edge positions were obtained via empirical correlations. Although our generative modeling framework was used to discover novel 2D photocatalysts for water splitting reaction, it is generic in nature and can be used directly to discover novel materials for other applications as well.

关键词： generative modeling variational autoencoder inverse design photocatalysts water splitting reaction

来源：评论

学校读者我要写书评

暂无评论

Recurrent variational autoencoders for Learning Nonlinear Generative Models in the Presence of Outliers

引用

IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING 2018年第6期12卷 1615-1627页

作者： Wang, Yu Dai, Bin Hua, Gang Aston, John Wipf, David Univ Cambridge Dept Pure Math & Stat Cambridge CB2 1TN England Tsinghua Univ Beijing 100084 Peoples R China Microsoft Res Redmond WA 98052 USA

This paper explores two useful modifications of the recent variational autoencoder (VAE), a popular deep generative modeling framework that dresses traditional autoencoders with probabilistic attire. The first involves a specially-tailored form of conditioning that allows us to simplify the VAE decoder structure while simultaneously introducing robustness to outliers. In a related vein, a second, complementary alteration is proposed to further build invariance to contaminated or dirty samples via a data augmentation process that amounts to recycling. In brief, to the extent that the VAE is legitimately a representative generative model, then each output from the decoder should closely resemble an authentic sample, which can then be resubmitted as a novel input ad infinitum. Moreover, this can be accomplished via special recurrent connections without the need for additional parameters to be trained. We evaluate these proposals on multiple practical outlier-removal and generative modeling tasks involving nonlinear low-dimensional manifolds, demonstrating considerable improvements over existing algorithms.

关键词： Deep generative models variational autoencoder robust PCA outlier removal variational Bayesian model deep learning

来源：评论

学校读者我要写书评

暂无评论

CROSS-VAE: TOWARDS DISENTANGLING EXPRESSION FROM IDENTITY FOR HUMAN FACES

CROSS-VAE: TOWARDS DISENTANGLING EXPRESSION FROM IDENTITY FO...

引用

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Wu, Haozhe Jia, Jia Xie, Lingxi Qi, Guojun Shi, Yuanchun Tian, Qi Tsinghua Univ Dept Comp Sci & Technol Beijing 100084 Peoples R China Beijing Natl Res Ctr Informat Sci & Technol BNRis Beijing Peoples R China Tsinghua Univ Inst Artificial Intelligence Beijing Peoples R China Huawei Noahs Ark Lab Montreal PQ Canada Futurewei Technol Plano TX USA

ISBN: (纸本)9781509066315

Facial expression and identity are two independent yet intertwined components for representing a face. For facial expression recognition, identity can contaminate the training procedure by providing tangled but irrelevant information. In this paper, we propose to learn clearly disentangled and discriminative features that are invariant of identities for expression recognition. However, such disentanglement normally requires annotations of both expression and identity on one large dataset, which is often unavailable. Our solution is to extend conditional VAE to a crossed version named Cross-VAE, which is able to use partially labeled data to disentangle expression from identity. We emphasis the following novel characteristics of our Cross-VAE: (1) It is based on an independent assumption that the two latent representations' distributions are orthogonal. This ensures both encoded representations to be disentangled and expressive. (2) It utilizes a symmetric training procedure where the output of each encoder is fed as the condition of the other. Thus two partially labeled sets can be jointly used. Extensive experiments show that our proposed method is capable of encoding expressive and disentangled features for facial expression. Compared with the baseline methods, our model shows an improvement of 3.56% on average in terms of accuracy.

关键词： Facial expression recognition Disentangle variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Stochastic Curiosity Exploration for Dialogue Systems 21

Stochastic Curiosity Exploration for Dialogue Systems

引用

Interspeech Conference

作者： Chien, Jen-Tzung Hsu, Po-Chien Natl Chiao Tung Univ Dept Elect & Comp Engn Hsinchu Taiwan

ISBN: (纸本)9781713820697

Traditionally, task-oriented dialogue system is built by an autonomous agent which can be trained by reinforcement learning where the reward from environment is maximized. The agent is learned by updating the policy when the goal state is observed. However, in real world, the extrinsic reward is usually sparse or missing. The training efficiency is bounded. The system performance is degraded. It is challenging to tackle the issue of sample efficiency in sparse reward scenario for spoken dialogues. Accordingly, a dialogue agent needs additional information to update its policy even in the period when reward is absent in the environment. This paper presents a new dialogue agent which is learned by incorporating the intrinsic reward based on the information-theoretic approach via stochastic curiosity exploration. This agent encourages the exploration for future diversity based on a latent dynamic architecture which consists of encoder network, curiosity network, information network and policy network. The latent states and actions are drawn to predict stochastic transition for future. The curiosity learning are implemented with intrinsic reward in a metric of mutual information and prediction error in the predicted states and actions. Experiments on dialogue management using PyDial demonstrate the benefit by using the stochastic curiosity exploration.

关键词： dialogue management deep reinforcement learning information-theoretic learning variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

On the Low-density Latent Regions of VAE-based Language Models

On the Low-density Latent Regions of VAE-based Language Mode...

引用

NeurIPS Workshop on Pre-Registration in Machine Learning

作者： Li, Ruizhe Peng, Xutan Lin, Chenghua Rong, Wenge Chen, Zhigang Univ Sheffield Dept Comp Sci Sheffield S Yorkshire England Beihang Univ Sch Comp Sci & Engn Beijing Peoples R China Zhejiang Wanli Univ Coll Informat & Intelligent Engn Ningbo Peoples R China

By representing semantics in latent spaces, variational autoencoders (VAEs) have been proven powerful in modelling and generating signals such as image and text, even without supervision. However, previous studies suggest that in a learned latent space, some low-density regions (aka. holes) exist, which could harm the overall system performance. While existing studies focus on empirically mitigating these latent holes, how they distribute and how they affect different components of a VAE, are still unexplored. In addition, the hole issue in VAEs for language processing is rarely addressed. In our work, by introducing a simple hole-detection algorithm based on the neighbour consistency between VAE's input, latent, and output semantic spaces, we propose to deeply dive into these topics for the first time. Comprehensive experiments including automatic evaluation and human evaluation imply that large-scale low-density latent holes may not exist in the latent space. In addition, various sentence encoding strategies are explored and the native word embedding is the most suitable strategy for VAEs in language modelling task.

关键词： variational autoencoder Low-density Regions Latent Holes

来源：评论

学校读者我要写书评

暂无评论

Learning Representations of Endoscopic Videos to Detect Tool Presence Without Supervision 10th

Learning Representations of Endoscopic Videos to Detect Tool...

引用

10th International Workshop on Multimodal Learning for Clinical Decision Support.(ML-CDS) / 9th International Workshop on Clinical Image-based Procedures - From Planning to Intervention (CLIP)

作者： Li, David Z. Ishii, Masaru Taylor, Russell H. Hager, Gregory D. Sinha, Ayushi Johns Hopkins Univ Dept Comp Sci Baltimore MD 21205 USA Johns Hopkins Med Inst Baltimore MD 21205 USA Johns Hopkins Univ Lab Computat Sensing & Robot Baltimore MD USA

ISBN: (纸本)9783030609450;9783030609467

In this work, we explore whether it is possible to learn representations of endoscopic video frames to perform tasks such as identifying surgical tool presence without supervision. We use a maximum mean discrepancy (MMD) variational autoencoder (VAE) to learn lowdimensional latent representations of endoscopic videos and manipulate these representations to distinguish frames containing tools from those without tools. We use three different methods to manipulate these latent representations in order to predict tool presence in each frame. Our fully unsupervised methods can identify whether endoscopic video frames contain tools with average precision of 71.56, 73.93, and 76.18, respectively, comparable to supervised methods. Our code is available at https:// ***/zdavidli/tool-presence/.

关键词： Endoscopic video Tool presence Representation learning variational autoencoder Maximum mean discrepancy

来源：评论

学校读者我要写书评

暂无评论

Simulating real profiles for shilling attacks: A generative approach

引用

KNOWLEDGE-BASED SYSTEMS 2021年 230卷 107390-107390页

作者： Barbieri, Julio Alvim, Leandro G. M. Braida, Filipe Zimbrao, Geraldo Univ Fed Rio de Janeiro PESC COPPE CT Cidade UnivPOB 68511 Rio De Janeiro Brazil Univ Fed Rural Rio de Janeiro DCC IM BR-26020740 Rio De Janeiro Brazil

Collaborative Filtering (CF) approaches are vulnerable to Shilling Attacks, in which malicious users or companies inject a large number of fake profiles in a system in order to manipulate its recommendations. One problem of current Shilling Attack models is that they commonly use straightforward statistical templates, producing profiles with different rating patterns than actual system data, which facilitates its detection, requiring a larger amount of profiles to achieve its goals. To address this problem and create profiles closer to reality, we propose using a generative model, variational autoencoder (VAE) to map original data distribution. With VAE, it is possible to generate new profiles based on real data, without explicit copying their actual ratings. Its generated profiles are converted to malicious profiles by adding target item rating value. We test our attack model on MovieLens 100k data set and compare to literature attack models. Our results indicate that our model outperforms all other models in model-based CF system, especially using low attack sizes (from 3% to 5%). Also, analysis comparing profiles generated from it and other approaches shows that our model ratings pattern are very similar to real profiles, which may indicate that attacks mounted using our approach may be less likely to be detected by detection approaches. Thus, we show that our attack model represents an advance on Shilling Attack models, since its superior results in model-based CF and possible indistinction from real profiles may be useful as a baseline to test detection techniques and other tasks among Shilling Attack area. (C) 2021 Elsevier B.V. All rights reserved.

关键词： Recommender systems Collaborative Filtering Shilling Attack Generative Model variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Multimodal Classification of Parkinson's Disease in Home Environments with Resiliency to Missing Modalities

引用

SENSORS 2021年第12期21卷 4133-4133页

作者： Heidarivincheh, Farnoosh McConville, Ryan Morgan, Catherine McNaney, Roisin Masullo, Alessandro Mirmehdi, Majid Whone, Alan L. Craddock, Ian Univ Bristol Sch Comp Sci Elect & Elect Engn & Engn Maths Bristol BS8 1UB Avon England Univ Bristol Med Sch Translat Hlth Sci Bristol BS8 1UD Avon England North Bristol NHS Trust Bristol Brain Ctr Bristol BS10 5PN Avon England Monash Univ Dept Human Centred Comp Melbourne Vic 3000 Australia

Parkinson's disease (PD) is a chronic neurodegenerative condition that affects a patient's everyday life. Authors have proposed that a machine learning and sensor-based approach that continuously monitors patients in naturalistic settings can provide constant evaluation of PD and objectively analyse its progression. In this paper, we make progress toward such PD evaluation by presenting a multimodal deep learning approach for discriminating between people with PD and without PD. Specifically, our proposed architecture, named MCPD-Net, uses two data modalities, acquired from vision and accelerometer sensors in a home environment to train variational autoencoder (VAE) models. These are modality-specific VAEs that predict effective representations of human movements to be fused and given to a classification module. During our end-to-end training, we minimise the difference between the latent spaces corresponding to the two data modalities. This makes our method capable of dealing with missing modalities during inference. We show that our proposed multimodal method outperforms unimodal and other multimodal approaches by an average increase in F1-score of 0.25 and 0.09, respectively, on a data set with real patients. We also show that our method still outperforms other approaches by an average increase in F1-score of 0.17 when a modality is missing during inference, demonstrating the benefit of training on multiple modalities.

关键词： Parkinson's disease deep learning multimodal data missing modality accelerometer computer vision variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：