Autoencoder & Natural Language 2015


Notes:

An autoencoder is a artificial neural network used for learning a compressed representation for feature extraction.

Resources:

  • nldb2015.org .. international conference on application of natural language to information systems

Wikipedia:

See also:

Autoencoder & Natural Language 2013 | Autoencoder & Natural Language 2014


A hierarchical neural autoencoder for paragraphs and documents J Li, MT Luong, D Jurafsky – arXiv preprint arXiv:1506.01057, 2015 – arxiv.org A Hierarchical Neural Autoencoder for Paragraphs and Documents … Natural language generation of coherent long texts like paragraphs or longer doc- uments is a challenging … step toward this generation task: training an LSTM (Long- short term memory) auto-encoder to pre … Cited by 48 Related articles All 15 versions

A primer on neural network models for natural language processing Y Goldberg – arXiv preprint arXiv:1510.00726, 2015 – arxiv.org … The tutorial covers input encoding for natural language tasks, feed-forward networks, convolutional … Other unsupervised approaches, including autoencoders and recursive au- toencoders, also fall out … also for CCG supertagging (Lewis & Steedman, 2014), dialog state tracking … Cited by 27 Related articles All 8 versions

[BOOK] Natural Language Processing and Information Systems: 20th International Conference on Applications of Natural Language to Information Systems, NLDB … C Biemann, S Handschuh, A Freitas, F Meziane… – 2015 – books.google.com … wide range of topics that all revolve around the use of natural language to access … and the lexicon, query processing, question answering, speech processing, and dialog systems … 153 Csaba Veres Comparing Recursive Autoencoder and Convolutional Network for Phrase-Level … Cited by 2 Related articles

Multi-domain dialogue success classifiers for policy training D Vandyke, PH Su, M Gasic, N Mrksic… – … IEEE Workshop on …, 2015 – ieeexplore.ieee.org … Using an autoencoder (AE) NN to perform dimen- sion reduction resulted in almost no loss … for handling the variable number of slots (eg recursive autoencoders [38], or … Systems: A Data-driven Methodology for Dialogue Management and Natural Language Generation, Springer … Cited by 4 Related articles All 8 versions

Deep learning approaches to problems in speech recognition, computational chemistry, and natural language text processing GE Dahl – 2015 – tspace.library.utoronto.ca … Similarly, in natural language processing, we can observe sequences of words with very little corruption, but the interactions between words that … specify an appropriate noise model (what is the noise model for the unobserved pragmatic context of human dialogue?); and we … Cited by 3 Related articles All 5 versions

Interpreting Questions with a Log-Linear Ranking Model in a Virtual Patient Dialogue System E Jaffe, M White, W Schuler… – Proceedings of the …, 2015 – anthology.aclweb.org … allows for easier authoring than, for example, systems that use deep natural language understanding (Dzikovska et … makes use of a notion of topic to organize the dialogue, which we … present another vector space method making use of recur- sive autoencoders, enabling vectors … Cited by 1 Related articles All 9 versions

An Auto-Encoder for Learning Conversation Representation Using LSTM X Zhou, B Hu, Q Chen, X Wang – International Conference on Neural …, 2015 – Springer … Additionally, we will integrate the interactive scheme into hierarchical auto-encoder [16] to explore … Li, J., Luong, M.-T., Jurafsky, D.: A hierarchical neural autoencoder for paragraphs … J., Huang, EH, Ng, AY, Manning, CD: Semi-supervised recursive autoencoders for predicting … Related articles

Natural Language Processing and Information Systems C Biemann, S Handschuh, A Freitas, F Meziane… – Springer … wide range of topics that all revolve around the use of natural language to access … and the lexicon, query processing, question answering, speech processing, and dialog systems … 153 Csaba Veres Comparing Recursive Autoencoder and Convolutional Network for Phrase-Level …

User Information Extraction for Personalized Dialogue Systems T Hirano, N Kobayashi, R Higashinaka, T Makino… – SEMDIAL 2015 …, 2015 – flov.gu.se … meaning by applying paraphrase detection meth- ods such as using recursive autoencoders (Socher et … In Proceed- ings of the 17th Annual Meeting of Association for Natural Language Processing (in … Acquisition and use of long-term memory for per- sonalized dialog systems. … Related articles All 5 versions

Topics, Trends, and Resources in Natural Language Processing (NLP) M Bansal – Citeseer Topics, Trends, and Resources in Natural Language Processing (NLP) Mohit Bansal TTI-Chicago … WSD, NER, Diachronics, Summarization, Generation, Multimodal, … Some Next Topics: Humor, Sarcasm, Idioms, Human-like Dialog, Poetry Page 7. Part-of-Speech Tagging … Related articles All 2 versions

3D object retrieval with stacked local convolutional autoencoder B Leng, S Guo, X Zhang, Z Xiong – Signal Processing, 2015 – Elsevier … capacity [23], and series of research related to deep learning have been boomed up in many fields such as 2-D image operating [24], speech recognition [25] and natural language processing [26]. … A stacked autoencoder just stacks several autoencoders? encoder part … Cited by 21 Related articles All 3 versions

Deep Markov Neural Network for Sequential Data Classification M Yang, W Tu, W Yin, Z Lu – Volume 2: Short Papers – aclweb.org … RAE: Recursive Autoencoder (Socher et al., 2011b) has been proven effective in many senti … Semi-Supervised Recursive Autoencoders for Pre- dicting Sentiment Distributions. … In Proceedings of the Conference on Em- pirical Methods in Natural Language Processing (EMNLP … Related articles All 7 versions

A diversity-promoting objective function for neural conversation models J Li, M Galley, C Brockett, J Gao, B Dolan – arXiv preprint arXiv: …, 2015 – arxiv.org … Earlier efforts to incorporate statistical methods into dialog systems typically relied on one of two ap- proaches. … 7IMSDB (http://www.imsdb.com/) is a relatively small database of around 0.4 million sentences and thus not suit- able for open domain dialogue training. … Cited by 33 Related articles All 5 versions

Deep Learning Techniques and its Various Algorithms and Techniques N Bhatia, MC Rana – 2015 – ijeir.org … dataset, it is better to use a batch method to train a sparse auto encoder because we … (d) Tiled and locally connected networks RBMs and auto encoders have densely … Natural Language Processing(NLP) is a typical example; deep learning cannot understand a story, as well as a … Related articles

Word embeddings combination and neural networks for robustness in asr error detection S Ghannay, Y Esteve, N Camelin – … (EUSIPCO), 2015 23rd …, 2015 – ieeexplore.ieee.org … approach to com- bine different word embeddings by using a denoising auto- encoder. … tation,” in Proceedings of the Empiricial Methods in Natural Language Processing (EMNLP … zagol, “Extracting and composing robust features with denoising autoencoders,” in Proceedings of … Cited by 5 Related articles All 3 versions

Leveraging User Ratings for Resource-poor Sentiment Classification NX Bach, TM Phuong – Procedia Computer Science, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. Help. … Semi-supervised Recursive Autoencoders for Predicting Sentiment Distri- butions. In: Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP) (2011), pp. … Cited by 3 Related articles All 2 versions

Closing the Gap: Domain Adaptation from Explicit to Implicit Discourse Relations YJGZJ Eisenstein – jiyfeng.github.io … from the marginalized denoising autoencoder for relation identification, we concatenate them with the original … 2012. Marginalized denoising autoencoders for do- main adaptation … In Proceedings of Empir- ical Methods for Natural Language Processing (EMNLP), pages 343– … Related articles All 10 versions

Towards a Model of Prediction-based Syntactic Category Acquisition: First Steps with Word Embeddings RGGCW Daelemans, S Gillis – … METHODS IN NATURAL LANGUAGE …, 2015 – aclweb.org … This is thought to facilitate understanding in dialogue. … given this line of work, that prediction is central to recently popular meth- ods from Natural Language Processing (NLP … 2.1 Memory Component: Auto Encoder During the first stage, we use a denoising Auto En- coder to (a … Related articles All 17 versions

Discourse Relation Recognition by Comparing Various Units of Sentence Expression with Recursive Neural Network A Otsuka, T Hirano, C Miyazaki, R Masumura… – 2015 – aclweb.org … 13th Annual Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL 2012 … Dy- namic Pooling and Unfolding Recursive Autoencoders for Paraphrase Detection. … Proc of the 1st Workshop on Natural Language and Automated Reasoning (NLPAR 2013), pages … Related articles All 8 versions

Robust deep-learning models for text-to-speech synthesis support on embedded devices T Boro?, SD Dumitrescu – … of the 7th International Conference on …, 2015 – dl.acm.org … The network itself is trained as follows: train iteratively each autoencoder layer, then fine-tune the … We run the network on 10 auto-encoder (AE) iterations (for each AE), with a further 2000 for … We use a 30 and 70 neurons auto encoders, and a 140 neuron output layer that should … Related articles

Sarcasm Detection in Social Media A Signhaniya, G Shenoy, R Kondekar – rohitkondekar.github.io … One promising approach in natural language understanding that has recently emerged (Mikolovet al., 2013) is … Use of Recursive Autoencoder (RAE) in its default settings does not yield a … in tweets would add to poorly trained word representations and hence the auto- encoder. … Related articles

Human Affect Recognition: Audio?Based Methods B Schuller, F Weninger – Wiley Encyclopedia of Electrical and …, 2015 – Wiley Online Library … For practical purposes, such as in dialog systems, one can choose sentence or subsentence units of analysis in accordance with the segmentation chosen for the acoustic features X. … Interoperability of emotion recognizers with existing systems, such as dialog systems, is crucial. … Cited by 1 Related articles

Negative Emotion Recognition in Spoken Dialogs X Zhang, H Wang, L Li, M Zhao, Q Li – … Linguistics and Natural Language …, 2015 – Springer … Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data. … Each layer is trained as a denoising autoencoder by minimizing the reconstruction of its … In our model, the denoising autoencoders can discover robust features for … Related articles All 2 versions

Exploiting Out-of-Domain Data Sources for Dialectal Arabic Statistical Machine Translation K Kirchhoff, B Zhao, W Wang – arXiv preprint arXiv:1509.01938, 2015 – arxiv.org … the evaluation system developed for a bilingual dialog system, and tested the system on … In Proceedings of the Second Workshop on Natural Language Pro- cessing for Social Media … 2014. An autoencoder with bilingual sparse features for improved statistical machine translation … Related articles All 3 versions

Syntax-based deep matching of short texts M Wang, Z Lu, H Li, Q Liu – arXiv preprint arXiv:1503.02427, 2015 – arxiv.org … In dialogue, the model needs to judge whether a message is an appropriate response to a given utterance. Deep neural network can model non-linear and hierar- chical relations [Bengio, 2009], and thus is well suited for short-text matching in natural language processing. The … Cited by 11 Related articles All 9 versions

Applying deep learning to answer selection: A study and an open task M Feng, B Xiang, MR Glass, L Wang… – 2015 IEEE Workshop …, 2015 – ieeexplore.ieee.org … Natural language understanding based spoken dialog system has been a popular topic in the past years of artificial intelli- gence renaissance. Many of those influential systems include a question answering module, eg Apple’s Siri, IBM’s Wat- son and Amazon’s Echo. … Cited by 18 Related articles All 4 versions

[BOOK] Advances in Artificial Intelligence: 28th Canadian Conference on Artificial Intelligence, Canadian AI 2015, Halifax, Nova Scotia, Canada, June 2-5, 2015, … D Barbosa, E Milios – 2015 – books.google.com … study of a type of opinion implicature (ie, opinion-oriented in- ference) in text and dialog. … the task, because deep learning was shown to work well for other natural language processing tasks … We found that stacked denoising auto-encoders are well suited for this task, with results …

Argumentation Mining in Parliamentary Discourse N Naderi, G Hirst – db.toronto.edu … Walker, MA: Using summarization to discover argu- ment facets in online idealogical dialog. … J., Huang, EH, Ng, AY, Manning, CD: Semi-supervised recur- sive autoencoders for predicting … In: Proceedings of the Conference on Empirical Methods in Natural Language Processing. … Cited by 3 Related articles All 6 versions

Structured Vectors for Chinese Word Representations C Li, B Xu, X Wang, G Wu, G Tian… – International Journal of …, 2015 – search.proquest.com … Huang, AY Ng, and CD Manning, Semi-supervised recursive auto encoders for predicting … His research interest includes natural language processing, big data, deep learning. … His current research interests include spoken dialogue systems, dialogue management, reinforcement … Related articles All 3 versions

Discriminative methods for statistical spoken dialogue systems MS Henderson – 2015 – repository.cam.ac.uk … 61 5.7.1 Denoising Autoencoder Initialisation . . . . . … Research on spoken dialogue systems seeks to create computer systems that can hold a conversation using natural language, just like a real human conversational partner. This is … The Dialog State Tracking Chal- … Cited by 4 Related articles All 3 versions

Tree recurrent neural networks with application to language modeling X Zhang, L Lu, M Lapata – arXiv preprint arXiv:1511.00060, 2015 – arxiv.org … related models. 1 Introduction Statistical language modeling plays an important role in many areas of natural language process- ing including speech recognition, machine trans- lation, and information retrieval. The prototyp … Cited by 1 Related articles All 3 versions

When are tree structures necessary for deep learning of representations? J Li, MT Luong, D Jurafsky, E Hovy – arXiv preprint arXiv:1503.00185, 2015 – arxiv.org Page 1. When Are Tree Structures Necessary for Deep Learning of Representations? Jiwei Li1, Minh-Thang Luong1, Dan Jurafsky1 and Eduard Hovy2 1Computer Science Department, Stanford University, Stanford, CA 94305 … Cited by 33 Related articles All 15 versions

Lcsts: A large scale chinese short text summarization dataset B Hu, Q Chen, F Zhu – arXiv preprint arXiv:1506.05865, 2015 – arxiv.org … Naturally Annotated Web Resources based Natural Language Processing is proposed by Sun (Sun, 2011 … Graves et al., 2013), machine transla- tion (Sutskever et al., 2014) and automatic dialog response (Shang … A hierarchical neural autoencoder for paragraphs and documents … Cited by 13 Related articles All 11 versions

Bilingual continuous-space language model growing for statistical machine translation R Wang, H Zhao, BL Lu, M Utiyama… – IEEE/ACM Transactions …, 2015 – ieeexplore.ieee.org … Siivola et al. present a method for estimating variable-length n-gram LM incrementally while maintaining some aspects of Kneser Ney smoothing [16], which is popular applied to several natural language processing tasks [41], [42], [43], [44], [45], [46], [47], [48]. … Cited by 10 Related articles All 7 versions

Softmax RNN for Short Text Classification E Wulczyn, C Jacoby – clementinejacoby.com … As a reminder in [3], Socher trains an unsupervised, recursive auto-encoder for sentence … A unified architecture for natural language processing: Deep neural networks with multitask learning. … Dy- namic pooling and unfolding recursive autoencoders for paraphrase detection. … Related articles

A critical review of recurrent neural networks for sequence learning ZC Lipton, J Berkowitz, C Elkan – arXiv preprint arXiv:1506.00019, 2015 – arxiv.org … Besides dialogue systems, modern interactive systems of economic importance include self-driving cars and robotic surgery … This is the case with natural language. … a stored pattern from a corrupted version and are the forerunners of Boltzmann machines and auto-encoders. 11 … Cited by 37 Related articles All 10 versions

Which ASR errors are hard to detect S Ghannay, N Camelin, Y Esteve – Errors by Humans and …, 2015 – errare2015.racai.ro … The denoising auto-encoder is composed of one hidden layer with 200 hidden units. … representation,” in Proceedings of the Empiri- cial Methods in Natural Language Processing (EMNLP … Extract- ing and composing robust features with denoising autoencoders,” in Proceedings … Cited by 2 Related articles All 2 versions

Advances in Artificial Intelligence D Barbosa, E Milios – Springer … study of a type of opinion implicature (ie, opinion-oriented in- ference) in text and dialog. … the task, because deep learning was shown to work well for other natural language processing tasks … We found that stacked denoising auto-encoders are well suited for this task, with results …

Syntax-based Deep Matching of Short Texts MWZLH Li, WJQ Liu – pdfs.semanticscholar.org … In dialogue, the model needs to judge whether a message is an appropriate response to a given utterance. Deep neural network can model non-linear and hierar- chical relations [Bengio, 2009], and thus is well suited for short-text matching in natural language processing. … Related articles All 2 versions

CroVeWA: Crosslingual Vector-Based Writing Assistance H Soyer, G Topic, P Stenetorp, A Aizawa – Proceedings of NAACL-HLT, 2015 – aclweb.org … 2014. An autoencoder approach to learning bilingual word representations. … 2008. A unified ar- chitecture for natural language processing: Deep neu- ral networks with multitask learning. … 2011. Semi- supervised recursive autoencoders for predicting sen- timent distributions. … Related articles All 7 versions

Towards universal paraphrastic sentence embeddings J Wieting, M Bansal, K Gimpel, K Livescu – arXiv preprint arXiv: …, 2015 – arxiv.org … Word embeddings have become ubiquitous in natural language processing (NLP … One approach is to train an autoencoder in an attempt to learn the latent structure of the sequence, whether it be a sentence with a parse tree (Socher et al., 2011), or a longer sequence such as a … Cited by 19 Related articles All 2 versions

Bringing machine learning and compositional semantics together P Liang, C Potts – Annu. Rev. Linguist., 2015 – annualreviews.org … 2007, Turney & Pantel 2010). The two types of approaches share the long-term vision of achieving deep natural language understanding, but their day-to-day differences can make them seem unrelated and even incompatible. … Cited by 19 Related articles All 4 versions

Video paragraph captioning using hierarchical recurrent neural networks H Yu, J Wang, Z Huang, Y Yang, W Xu – arXiv preprint arXiv:1510.07712, 2015 – arxiv.org … ing human activities with natural language [23]. … Applying RNN to translating visual sequence to natural lan- guage is largely inspired by the recent advances in Neural Machine Translation (NMT) [1, 42] in the natural language processing community. … Cited by 12 Related articles All 5 versions

Adequacy–fluency metrics: Evaluating MT in the continuous space model framework RE Banchs, LF D’Haro, H Li – IEEE/ACM Transactions on Audio, …, 2015 – ieeexplore.ieee.org … Index Terms—System Evaluation, Machine Translation, Natural Language Processing. I. INTRODUCTION … meaning has been proven to be useful in many monolingual and cross-language natural language processing applications such as … Cited by 8 Related articles All 3 versions

Transfer learning using computational intelligence: a survey J Lu, V Behbood, P Hao, H Zuo, S Xue… – Knowledge-Based …, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect … features at different levels, such as word level and sentence level in Natural Language Processing … The Stacked Denoising Autoencoder (SDA) is another structure that is presented in deep neural network [50 … Cited by 42 Related articles All 5 versions

Learning deep representations via extreme learning machines W Yu, F Zhuang, Q He, Z Shi – Neurocomputing, 2015 – Elsevier … Forgotten username or password? Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. Export file. … The performances of the deep models (ie Stacked Auto-encoder) are comparable. … Cited by 15 Related articles All 4 versions

Leveraging valence and activation information via multi-task learning for categorical emotion recognition R Xia, Y Liu – 2015 IEEE International Conference on Acoustics …, 2015 – ieeexplore.ieee.org … In [12], the authors modified the auto-encoders and used two hidden representations (one … task learning has been widely applied to many speech and natural language processing related … Yang Liu, and Bjorn Schuller, “In troducing shared-hidden-layer autoencoders for transfer … Cited by 2 Related articles

Deep correspondence restricted Boltzmann machine for cross-modal retrieval F Feng, R Li, X Wang – Neurocomputing, 2015 – Elsevier … Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. … Very recently, two effective one-stage methods, multi-modal stacked autoencoders (MSAE) [9] and correspondence autoencoder (Corr-AE … Cited by 7 Related articles All 3 versions

Semi-automatic Filtering of Translation Errors in Triangle Corpus SK Choi, JH Shin, YG Kim – 2015 – anthology.aclweb.org … This evaluation set has same tourist/dialog domains as crowdsourcing translation corpus. … In Proceedings of the 5th International Joint Conference on Natural Language Processing, 1361 … H. (2014) Learning multilingual word representations using a bag-of-words autoencoder. … Related articles All 8 versions

Towards Universal Paraphrastic Sentence Embeddings JWMBK Gimpel, K Livescu – arXiv preprint arXiv: …, 2015 – pdfs.semanticscholar.org … Word embeddings have become ubiquitous in natural language processing (NLP … One approach is to train an autoencoder in an attempt to learn the latent structure of the sequence, whether it be a sentence with a parse tree (Socher et al., 2011), or a longer sequence such as a … Related articles All 4 versions

Boosting paraphrase detection through textual similarity metrics with abductive networks ESM El-Alfy, RE Abdel-Aal, WG Al-Khatib, F Alvi – Applied Soft Computing, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … The vast amount of information available through electronic media and over the Internet has motivated several research works in natural language processing, machine translation and information retrieval. … Cited by 4 Related articles All 3 versions

Identifying synonymy between relational phrases using word embeddings NTH Nguyen, M Miwa, Y Tsuruoka, S Tojo – Journal of biomedical …, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. Help. … domain benefit from automatic clustering of relational phrases into synonymous groups, since it alleviates the problem of spurious mismatches caused by the diversity of natural language expressions. … Related articles All 8 versions

Cross-language acoustic emotion recognition: An overview and some tendencies SM Feraru, D Schuller – Affective Computing and Intelligent …, 2015 – ieeexplore.ieee.org Page 1. Cross-Language Acoustic Emotion Recognition: An Overview and Some Tendencies Silvia Monica Feraru1, Dagmar Schuller2, and Björn Schuller1,2,3,4 1Machine Intelligence & Signal Processing group, MMK, Technische … Cited by 5 Related articles All 4 versions

Negative Emotion Recognition in Spoken Dialogs Q Li – … Computational Linguistics and Natural Language …, 2015 – books.google.com … Each layer is trained as a denoising autoencoder by minimizing the reconstruction of its input … In our model, the denoising autoencoders can discover robust features for classification so as to … from a call center, where actual customers are engaged in spoken dialog with human … Related articles

The DIRHA-ENGLISH corpus and related tasks for distant-speech recognition in domestic environments M Ravanelli, L Cristoforetti, R Gretter… – … IEEE Workshop on …, 2015 – ieeexplore.ieee.org … while music/speech prompts are played), speaker verification, con- current dialogue management (to … T. Kawahara, “Reverberant speech recognition combining deep neural networks and deep autoencoders,” in Proc. … of the Workshop on Speech and Natural Language, 1992, pp … Cited by 4 Related articles

Minimally-Constrained Multilingual Embeddings via Artificial Code-Switching M Wick, P Kanani, A Pocock – 2015 – aaai.org … Introduction An important practical problem in natural language process- ing (NLP) is to make NLP tools (eg … and code switching in which a multilin- gual speaker switches between languages during a dialogue (Lipski 1978 … 2014), regularized auto-encoders (Sarath Chandar et al … Cited by 2 Related articles All 5 versions

A Multifaceted Approach to Sentence Similarity HT Nguyen, PH Duong, TQ Le – International Symposium on Integrated …, 2015 – Springer … 182–190 (2012) 8. Socher, R., Huang, EH, Pennington, J., Ng, AY, Manning, CD: Dynamic pooling and unfolding recursive autoencoders for paraphrase … In: Recent Advances in Natural Language Processing V, vol. … In: Current and new directions in discourse and dialogue, pp. … Cited by 3 Related articles All 3 versions

Learning to Grade Short Answers using Machine Learning Techniques R Krithika, J Narayanan – … of the Third International Symposium on …, 2015 – dl.acm.org … based learning; Information extraction; Phonology / mor- phology; Language resources; Vagueness and fuzzy logic; On- tology engineering; Neural networks; Discourse, dialogue and pragmatics … The broader aim is to be able to understand any natural language answer. … Cited by 2 Related articles

3D object understanding with 3D Convolutional Neural Networks B Leng, Y Liu, K Yu, X Zhang, Z Xiong – Information Sciences, 2015 – Elsevier … 3], deep Boltzmann machines (DBM) [64] and [65], and stacked autoencoder [69] and [70 … such as 2D image processing [35], speech recognition [13] and natural language processing [12]. … one convolutional layer, we consider the layer as a convolutional auto-encoder [54], which … Cited by 2 Related articles

Comparing attribute classifiers for interactive language grounding Y Yu, A Eshghi, O Lemon – pdfs.semanticscholar.org … at- tributes using L2-loss linear SVMs and to learn the associations between visual attributes and particu- lar words using Auto-encoders. … Our goal is to couple attribute classifiers with much wider coverage to the formal semantics of a full Natural Language dialogue system. … Cited by 3 Related articles All 7 versions

Feature learning based on SAE–PCA network for human gesture recognition in RGBD images SZ Li, B Yu, W Wu, SZ Su, RR Ji – Neurocomputing, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … In this paper, a feature learning approach based on sparse auto-encoder (SAE) and principle component analysis is proposed for recognizing human actions, ie finger-spelling or sign language, for RGB-D … Cited by 13 Related articles All 2 versions

Structural information aware deep semi-supervised recurrent neural network for sentiment analysis W Rong, B Peng, Y Ouyang, C Li, Z Xiong – Frontiers of Computer Science, 2015 – Springer … Currently a lot of approaches from natural language processing’s perspec- tive have been employed to conduct this task. … 2.3 Word embedding Embedding words into a continuous vector space has a long history in the domain of natural language processing. Ben- gio et al. … Cited by 3 Related articles All 4 versions

Sparse auto-encoder based feature learning for human body detection in depth image SZ Su, ZH Liu, SP Xu, SZ Li, R Ji – Signal Processing, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … Sparse Auto-Encoder (SAE) is an unsupervised feature leaning methods which can avoid the … Experiments on various applications are encouraging, such as natural language process, computer vision, and … Cited by 7 Related articles All 3 versions

A comparative study on selecting acoustic modeling units in deep neural networks based large vocabulary Chinese speech recognition X Li, Y Yang, Z Pang, X Wu – Neurocomputing, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … further research indicated that they could be trained in many different ways, such as the discriminative pre-training method introduced in [9], and generative pre-training with various types of auto-encoder. … Cited by 4 Related articles All 3 versions

Trends in extreme learning machines: a review G Huang, GB Huang, S Song, K You – Neural Networks, 2015 – Elsevier … Username: Password: Remember me. | Not Registered? Forgotten username or password? Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. Export file. … Cited by 198 Related articles All 9 versions

Machine learning for dialog state tracking: A review M Henderson – 2015 – research.google.com … not only from errors in recognizing speech, but also from ambiguities inherent in natural language, and so it is … the feature vectors are initial- ized with weights learned by training a denoising auto-encoder. … [4] JD Williams, A. Raux, and M. Henderson, “The Dialog State Tracking … Cited by 6 Related articles All 2 versions

No-reference image quality assessment with shearlet transform and deep neural networks Y Li, LM Po, X Xu, L Feng, F Yuan, CH Cheung… – Neurocomputing, 2015 – Elsevier … of attention and achieved great success on various applications, such as denoising [14] and [15], inpainting [15], classification [16] and natural language processing [17]. … A stacked autoencoder is a neural network consisting of multiple layers of sparse autoencoders in which … Cited by 9 Related articles All 3 versions

IODA: an Input/Output Deep Architecture for image labeling J Lerouge, R Herault, C Chatelain, F Jardin… – Pattern Recognition, 2015 – Elsevier … of p(y|X) p ( y | X ) . These approaches have shown to be efficient on numerous problems such as natural language processing [24 … learning of a DNN, an unsupervised pre-training is performed on deepest layers, through the use of auto-encoders (AE) which … 3.1.2. Auto-encoder. … Cited by 6 Related articles All 8 versions

Visual analysis of online social media to open up the investigation of stance phenomena K Kucher, T Schamp-Bjerede, A Kerren… – Information …, 2015 – ivi.sagepub.com … in automated ways of text processing that can be offered by researchers from the field of compu- tational linguistics or natural language processing (NLP). … is to convey the speaker’s viewpoint of what is talked about and to reg- ulate the exchange between the dialog partners. … Cited by 4 Related articles All 10 versions

Real-Time Topic and Sentiment Analysis in Human-Robot Conversation E Russell – 2015 – epublications.marquette.edu … the system build that supports these processes. A sample dialogue with the … 2.2 Sentiment Analysis Sentiment Analysis is a natural language processing sub-field that is … using recursive auto-encoders [38]. Unfortunately, most applications are extremely domain-dependent; it can … Related articles All 2 versions

Human Category Learning: Toward a Broader Explanatory Account KJ Kurtz – Psychology of Learning and Motivation, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … The DIVA model (Kurtz, 2007), uses a divergent autoencoder network architecture to instantiate the key … or multiclass classifications can be learned generatively by training coordinated autoencoders on the … Cited by 1 Related articles All 2 versions

Effectively classifying short texts by structured sparse representation with dictionary filtering L Gao, S Zhou, J Guan – Information Sciences, 2015 – Elsevier … Username: Password: Remember me. | Not Registered? Forgotten username or password? Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. Export file. … Cited by 1 Related articles All 3 versions

Predicting primary categories of business listings for local search ranking C Khan, J Lee, R Blanco, Y Chang – Neurocomputing, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … been a blooming field of interest in the machine learning and natural language processing/information … hierarchy and propose a method that learns compact class codes using autoencoders which leverage … Cited by 1 Related articles All 4 versions

A Multi-task Learning Framework for Emotion Recognition Using 2D Continuous Space R Xia, Y Liu – ieeexplore.ieee.org … In [17], the authors modified the auto-encoders to use two hidden representations (one for … task learning has been widely applied to many speech and natural language processing related … 5 female) acting in two different scenarios: scripted play and spontaneous dialog, in their … Cited by 2 Related articles All 2 versions

Multi-ganglion ANN based feature learning with application to P300-BCI signal classification W Gao, J Guan, J Gao, D Zhou – Biomedical Signal Processing and Control, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … proposed some other typical neural network for unsupervised learning, such as auto-encoder [26] and … These methods have been applied to image processing and natural language processing [28] and [29 … Cited by 2 Related articles All 2 versions

Scene classification based on single-layer SAE and SVM H Yin, X Jiao, Y Chai, B Fang – Expert Systems with Applications, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … In this paper, a scene classification approach based on single-layer sparse autoencoder (SAE) and … & Schuller, 2013), digit image recognition (Hinton et al., 2006), and natural language processing(Bengio … Cited by 13 Related articles All 3 versions

Learning to understand phrases by embedding the dictionary F Hill, K Cho, A Korhonen, Y Bengio – arXiv preprint arXiv:1504.00548, 2015 – arxiv.org … 2 Neural Language Model Architectures The first model we apply to the dictionary-based learning task is a recurrent neural network (RNN). RNNs operate on variable-length sequences of in- puts; in our case, natural language definitions, descriptions or sentences. … Cited by 13 Related articles All 10 versions

Detection and reconstruction of clipped speech for speaker recognition F Bie, D Wang, J Wang, TF Zheng – Speech Communication, 2015 – Elsevier … Username: Password: Remember me. | Not Registered? Forgotten username or password? Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. Export file. … Cited by 1 Related articles All 9 versions

Imaging-based enrichment criteria using deep learning algorithms for efficient clinical trials in mild cognitive impairment VK Ithapu, V Singh, OC Okonkwo, RJ Chappell… – Alzheimer’s & …, 2015 – Elsevier … This new disease marker (which we refer to as randomized denoising autoencoder marker, rDAm) is a machine learning … ideas in “deep learning” that yield state-of-the-art results in computer vision, natural language processing, and … 2.1.1. Randomized denoising autoencoders. … Cited by 3 Related articles All 8 versions

Subset based deep learning for RGB-D object recognition J Bai, Y Wu, J Zhang, F Chen – Neurocomputing, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … Then a RGB-Subset- Sparse auto-encoder was trained to extract features from RGB images and a Depth-Subset-Sparse auto-encoder was trained to extract features from depth images for each subset. … Cited by 13 Related articles All 3 versions

Promises and challenges of big data computing in health sciences T Huang, L Lan, X Fang, P An, J Min, F Wang – Big Data Research, 2015 – Elsevier … Username: Password: Remember me. | Not Registered? Forgotten username or password? Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. Export file. … Cited by 24 Related articles All 2 versions

Comparative Study of Sentiment Detection Techniques for Business Analytics H Avery – 2015 – etd.auburn.edu Page 1. Comparative Study of Sentiment Detection Techniques for Business Analytics by Heather Avery A dissertation submitted to the Graduate Faculty of Auburn University in partial fulfillment of the requirements for the Degree … Related articles All 3 versions

A knowledge-based multi-layered image annotation system M Ivasic-Kos, I Ipsic, S Ribaric – Expert Systems With Applications, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. … In (Yin, Jiao, Chai, & Fang, 2015) discriminant scene features were learned using single-layer sparse autoencoder (SAE) and then SVM classifier is used for scene classification. … Cited by 3 Related articles All 7 versions

An unsupervised feature learning framework for basal cell carcinoma image analysis J Arevalo, A Cruz-Roa, V Arias, E Romero… – Artificial intelligence in …, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. Help. … UFL) framework for histopathology image analysis that comprises three main stages: (1) local (patch) representation learning using different strategies (sparse autoencoders, reconstruct independent … Cited by 11 Related articles All 7 versions

Machine learning for gesture recognition from videos BG Gebre – 2015 – pubman.mpdl.mpg.de … I thank Marcos for keeping me interested in natural language processing (text processing … learning algorithms available (eg au- toencoders, clustering, dictionary learning, restricted Boltzmann machines), we implemented clustering (K-means) and sparse autoencoder algorithms … Related articles All 2 versions

Recurrent Neural Networks in Speech Disfluency Detection and Punctuation Prediction M Reisser – 2015 – isl.anthropomatik.kit.edu … However, designing automated systems that understand spoken language and are able to perform natural language processing (NLP) tasks such as automated translation, information extraction or dialogue systems, is a very challenging task. …

Image automatic annotation via multi-view deep representation Y Yang, W Zhang, Y Xie – Journal of Visual Communication and Image …, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … W 4 W 4 . Generally, the deep neural networks with pretraining by AE are also called as Stacked Auto-Encoder (SAE). Because the deep neural network in this model is stacked with several Auto-Encoders. … Related articles All 3 versions

Deep Neural Networks in Speech Recognition AL Maas – 2015 – stacks.stanford.edu … 12 2.2 Deep Recurrent Denoising Autoencoder. The network has two hidden … during an understanding task. Following transcription the system must perform some sort of natural language … tion process. We introduce a model which uses a deep recurrent auto encoder … Related articles

Electronic noses for food quality: A review A Loutfi, S Coradeschi, GK Mani, P Shankar… – Journal of Food …, 2015 – Elsevier … Article suggestions will be shown in a dialog on return to ScienceDirect. … Other approaches have attempted to provide human like descriptors to an electronic nose response using natural language symbols which have been manipulated to provide new descriptions to unseen … Cited by 61 Related articles All 7 versions

Automated Analysis of L2 French Writing: a preliminary study NL Parslow – 2015 – researchgate.net … our eons of practice, significant improvements in methods of acquisition remain elusive. With the rapid advances in Natural Language Processing (NLP) of recent years, the field of Computer Assisted Language Learning (CALL) has begun to integrate more and more …

Tensor representation learning based image patch analysis for text identification and recognition G Zhong, M Cheriet – Pattern Recognition, 2015 – Elsevier … Username: Password: Remember me. | Not Registered? Forgotten username or password? Help. Download PDF Opens in a new window. Article suggestions will be shown in a dialog on return to ScienceDirect. Help. Direct export. Export file. … Cited by 4 Related articles All 4 versions

[BOOK] Video Cataloguing: Structure Parsing and Content Extraction G Gao, CH Liu – 2015 – books.google.com … Meanwhile, a video story is some type of natural language description. … Meanwhile, although the video story description can better meet people’s watching or consumption demand, it is directly derived from the angle of human natural language. … Related articles All 5 versions

Compositional Distributional Semantics with Compact Closed Categories and Frobenius Algebras D Kartsaklis – arXiv preprint arXiv:1505.00138, 2015 – arxiv.org … Compositionality in semantics offers an elegant way to address the inherent property of natural language to produce an infinite number of structures (phrases and sen- … of it can be spotted even in works of Plato. In his dialogue Sophist, Plato argues … Cited by 15 Related articles All 6 versions

Deep neural networks for single-channel multi-talker speech recognition C Weng, D Yu, ML Seltzer… – IEEE/ACM Transactions …, 2015 – ieeexplore.ieee.org Page 1. 2329-9290 (c) 2015 IEEE. Personal use is permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org/ publications_standards/publications/rights/index.html for more information. This … Cited by 6 Related articles All 4 versions

Dynamic topic adaptation for improved contextual modelling in statistical machine translation EC Hasler – 2015 – era.lib.ed.ac.uk … machine translation. The ongoing efforts to produce high-quality parallel corpora as well as the vastly growing amounts of data on the web mean that more and more natural language data is available for training translation systems. On the one hand, … Cited by 1 Related articles All 5 versions

Concit-Corpus: Context Citation Analysis to learn Function, Polarity and Influence M Hernández Álvarez, JM Gómez Soriano – 2015 – rua.ua.es Page 1. Concit-Corpus: Context Citation Analysis to learn Function, Polarity and Influence Myriam Hernández-Álvarez Page 2. TESIS DOCTORAL Septiembre, 2015 Concit-Corpus: Context Citation Analysis to learn Function, Polarity and Influence Myriam Hernández-Álvarez … Related articles All 3 versions

Authorship verification A Stolerman – 2015 – idea.library.drexel.edu … 108]. This example is but an illustration of how stylometry has become dominated by computational methods in the last decades, specifically artificial intelligence applications involving natural language processing for quantifying … Cited by 1 Related articles All 2 versions

Adjoining Chaos Q Wang – … Methods in Computational Science, Engineering, and …, 2015 – drops.dagstuhl.de Page 30. 28 14371–Adjoint Methods in Computational Science, Engineering, and Finance 3.31 Adjoining Chaos Qiqi Wang (MIT–Cambridge, US) License Creative Commons BY 3.0 Unported license © Qiqi Wang Joint work … Related articles All 8 versions