DNN (Deep Neural Network) & Human Language Technology 2017


Notes:

Unlike lexical semantics, which focuses on the meanings of individual words, the field of compositional semantics looks at the meanings of sentences and longer utterances.

Resources:

Wikipedia:

References:

See also:

100 Best Deep Belief Network Videos100 Best Deep Learning Videos100 Best DeepMind Videos100 Best GitHub: Deep Learning100 Best Natural Language Deep Learning VideosDeep Belief Network & Dialog SystemsDeep Inference 2017Deep Reasoning & Dialog SystemsDeepDiveDNLP (Deep Natural Language Processing) | Skipgrams, Deep Learning & Question Answering 2017


Speaker diarization using deep neural network embeddings
D Garcia-Romero, D Snyder, G Sell… – … , Speech and Signal …, 2017 – ieeexplore.ieee.org
SPEAKER DIARIZATION USING DEEP NEURAL NETWORK EMBEDDINGS Daniel Garcia-Romero, David Snyder, Gregory Sell, Daniel Povey, and Alan McCree Human Language Technology Center of Excellence & Center for Language and Speech Processing The Johns …

Advances in all-neural speech recognition
G Zweig, C Yu, J Droppo… – Acoustics, Speech and …, 2017 – ieeexplore.ieee.org
… [3] F. Seide, G. Li, and D. Yu, “Conversational speech transcription using context-dependent deep neural networks”, in Interspeech … Woodland, “Tree-based state tying for high accuracy acoustic modelling”, in Proceedings of the work- shop on Human Language Technology, pp …

Visual genome: Connecting language and vision using crowdsourced dense image annotations
R Krishna, Y Zhu, O Groth, J Johnson, K Hata… – International Journal of …, 2017 – Springer
… processing. Syntactic features (Zhou et al. 2007; GuoDong et al. 2005), dependency tree methods (Culotta and Sorensen 2004; Bunescu and Mooney 2005), and deep neural networks (Socher et al. 2012; Zeng et al. 2014) have …

Deep neural machine translation with linear associative unit
M Wang, Z Lu, J Zhou, Q Liu – arXiv preprint arXiv:1705.00861, 2017 – arxiv.org
… Abstract Deep Neural Networks (DNNs) have provably enhanced the state-of-the- art Neural Machine Translation (NMT) with their capability in modeling com- plex functions and capturing com- plex linguistic structures. However …

A comprehensive study of deep bidirectional LSTM RNNs for acoustic modeling in speech recognition
A Zeyer, P Doetsch, P Voigtlaender… – … , Speech and Signal …, 2017 – ieeexplore.ieee.org
… Human Language Technology and Pattern Recognition, Computer Science Department, RWTH Aachen University, 52062 Aachen, Germany {zeyer, doetsch … 1. INTRODUCTION AND RELATED WORK Deep neural networks (DNN) yield state-of-the-art perfor- mance in …

Reporting score distributions makes a difference: Performance study of lstm-networks for sequence tagging
N Reimers, I Gurevych – arXiv preprint arXiv:1707.09861, 2017 – arxiv.org
… In recent years, deep neural networks were shown to achieve state-of-the-art performance for a wide range of NLP tasks, including many sequence tag- ging tasks (Ma and Hovy, 2016), dependency pars- ing (Andor et al., 2016), and machine translation (Wu et al., 2016) …

Sentiment analysis leveraging emotions and word embeddings
M Giatsoglou, MG Vozalis, K Diamantaras… – Expert Systems with …, 2017 – Elsevier
Skip to main content …

Improving sentiment analysis via sentence type classification using BiLSTM-CRF and CNN
T Chen, R Xu, Y He, X Wang – Expert Systems with Applications, 2017 – Elsevier
… achieves state-of-the-art results on several benchmarking datasets. Keywords. Natural language processing. Sentiment analysis. Deep neural network. 1. Introduction. Sentiment analysis is the field of study that analyzes people’s …

Recent trends in deep learning based natural language processing
T Young, D Hazarika, S Poria, E Cambria – arXiv preprint arXiv …, 2017 – arxiv.org
Page 1. 1 Recent Trends in Deep Learning Based Natural Language Processing Tom Younga†, Devamanyu Hazarikab†, Soujanya Poriac†, Erik Cambriad? a School of Information and Electronics, Beijing Institute of Technology …

Deep Neural Network Embeddings for Text-Independent Speaker Verification
D Snyder, D Garcia-Romero, D Povey… – Proc. Interspeech …, 2017 – pdfs.semanticscholar.org
Deep Neural Network Embeddings for Text-Independent Speaker Verification David Snyder, Daniel Garcia-Romero, Daniel Povey, Sanjeev Khudanpur Center for Language and Speech Processing & Human Language Technology Center of Excellence, The Johns Hopkins …

A morphology-aware network for morphological disambiguation
E Yildiz, C Tirkaz, HB Sahin, MT Eren… – arXiv preprint arXiv …, 2017 – arxiv.org
… A unified architecture for natural language processing: Deep neural networks with multitask learning … In Proceedings of the Main Conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics, HLT …

Adversarial generation of natural language
S Rajeswar, S Subramanian, F Dutil, C Pal… – arXiv preprint arXiv …, 2017 – arxiv.org
… A conditional version is also described that can generate sequences con- ditioned on sentence characteristics. 1 Introduction Deep neural networks have recently enjoyed some success at modeling natural language (Mikolov et al., 2010; Zaremba et al., 2014; Kim et al., 2015) …

Paraphrasing revisited with neural machine translation
J Mallinson, R Sennrich, M Lapata – … of the 15th Conference of the …, 2017 – aclweb.org
… Luong et al., 2015). At its core, NMT uses a deep neural network trained end-to-end to maximize the conditional probabil- ity of a correct translation given a source sen- tence, using a bilingual corpus. NMT models have obtained …

A first look into a Convolutional Neural Network for speech emotion detection
D Bertero, P Fung – Acoustics, Speech and Signal Processing …, 2017 – ieeexplore.ieee.org
… Human Language Technology Center Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology, Clear Water … 2]. In recent years people have delegated the role of learning emotional models to deep neural networks, which have …

An analysis of convolutional long short-term memory recurrent neural networks for gesture recognition
E Tsironi, P Barros, C Weber, S Wermter – Neurocomputing, 2017 – Elsevier
Skip to main content …

Toward controlled generation of text
Z Hu, Z Yang, X Liang… – International …, 2017 – proceedings.mlr.press
Page 1. Toward Controlled Generation of Text Zhiting Hu 1 2 Zichao Yang 1 Xiaodan Liang 1 2 Ruslan Salakhutdinov 1 Eric P. Xing 1 2 Abstract Generic generation and manipulation of text is challenging and has limited success …

Semi-supervised multitask learning for sequence labeling
M Rei – arXiv preprint arXiv:1704.07156, 2017 – arxiv.org
Page 1. Semi-supervised Multitask Learning for Sequence Labeling Marek Rei The ALTA Institute Computer Laboratory University of Cambridge United Kingdom marek.rei@cl.cam.ac.uk Abstract We propose a sequence labeling …

Label-dependencies aware recurrent neural networks
Y Dupont, M Dinarelli, I Tellier – arXiv preprint arXiv:1706.01740, 2017 – arxiv.org
Page 1. Label-Dependencies Aware Recurrent Neural Networks Yoann Dupont, Marco Dinarelli, Isabelle Tellier LaTTiCe (UMR 8094), CNRS, ENS Paris, Université Sorbonne Nouvelle – Paris 3 PSL Research University, USPC …

A Framework for pre-training hidden-unit conditional random fields and its extension to long short term memory networks
YB Kim, K Stratos, R Sarikaya – Computer Speech & Language, 2017 – Elsevier
Skip to main content …

The technology behind personal digital assistants: an overview of the system architecture and key components
R Sarikaya – IEEE Signal Processing Magazine, 2017 – ieeexplore.ieee.org
Page 1. 67 IEEE SIgnal ProcESSIng MagazInE | January 2017 | 1053-5888/17©2017IEEE e have long envisioned that one day computers will understand natural language and anticipate what we need, when and where we need it, and proactively complete tasks on our behalf …

Neural End-to-End Learning for Computational Argumentation Mining
S Eger, J Daxenberger, I Gurevych – arXiv preprint arXiv:1704.06104, 2017 – arxiv.org
Page 1. arXiv:1704.06104v1 [cs.CL] 20 Apr 2017 Neural End-to-End Learning for Computational Argumentation Mining Steffen Eger†‡, Johannes Daxenberger†, Iryna Gurevych†‡ †Ubiquitous Knowledge Processing Lab (UKP …

Re-Sign: Re-Aligned End-to-End Sequence Modelling with Deep Recurrent CNN-HMMs
O Koller, S Zargaran, H Ney – IEEE Conference on …, 2017 – openaccess.thecvf.com
… Oscar Koller, Sepehr Zargaran and Hermann Ney Human Language Technology & Pattern Recognition Group RWTH Aachen University, Germany {koller … can be observed performing re-alignments with GMM- free systems [33] purely based on deep neural networks, which is …

Modeling under-resourced languages for speech recognition
M Kurimo, S Enarvi, O Tilk, M Varjokallio… – Language Resources …, 2017 – Springer
… Estonian acoustic models are hybrid deep neural networks based hidden Markov models (DNN-HMMs) that use speaker identity vectors (i-vectors) as additional input features to the DNNs in parallel with the regular acoustic features, thus performing unsupervised transcript-free …

Deep learning for extracting protein-protein interactions from biomedical literature
Y Peng, Z Lu – arXiv preprint arXiv:1706.01556, 2017 – arxiv.org
Page 1. Deep learning for extracting protein-protein interactions from biomedical literature Yifan Peng Zhiyong Lu National Center for Biotechnology Information National Library of Medicine National Institutes of Health Bethesda, MD 20894 1yifan.peng, zhiyong.lul@nih.gov …

Generative encoder-decoder models for task-oriented spoken dialog systems with chatting capability
T Zhao, A Lu, K Lee, M Eskenazi – arXiv preprint arXiv:1706.08476, 2017 – arxiv.org
… Le, 2015; Sordoni et al., 2015). It encodes the dialog history using deep neural networks and then generates the next sys- tem utterance word-by-word via recurrent neural networks (RNNs). Therefore, unlike the tradi- tional …

Machine translation evaluation with neural networks
F Guzmán, S Joty, L Màrquez, P Nakov – Computer Speech & Language, 2017 – Elsevier
… Keywords. Machine translation. Reference-based MT evaluation. Deep neural networks. Distributed representation of texts. Textual similarity. 1. Introduction. Automatic machine translation (MT) evaluation is a necessary step when developing or comparing MT systems …

A comparative review of dynamic neural networks and hidden Markov model methods for mobile on-device speech recognition
MK Mustafa, T Allen, K Appiah – Neural Computing and Applications, 2017 – Springer
… However, these experiments were conducted on a computer, rather than mobile phone, as initial test runs of the deep neural network experiments were found to be computa- tionally intensive even on a … In: Proceedings of the workshop on Human Language Technology …

Distant Supervision for Relation Extraction with Sentence-Level Attention and Entity Descriptions.
G Ji, K Liu, S He, J Zhao – AAAI, 2017 – aaai.org
Page 1. Distant Supervision for Relation Extraction with Sentence-Level Attention and Entity Descriptions Guoliang Ji, Kang Liu, Shizhu He, Jun Zhao National Laboratory of Pattern Recognition (NLPR) Institute of Automation …

Improved subword modeling for WFST-based speech recognition
P Smit, S Virpioja, M Kurimo – Proc. INTERSPEECH, 2017 – research.aalto.fi
… The acoustic models are all sequence-trained deep neural network models [23] trained with the Kaldi toolkit [4]. This … M. Saraclar, “Unlimited vocabulary speech recognition for agglutinative languages,” in Proceedings of the 2006 Human Language Technology Conference of …

Using coreference links to improve spanish-to-english machine translation
LM Werlen, A Popescu-Belis – Proceedings of the 2nd Workshop on …, 2017 – aclweb.org
… (2015) designed a classifier based on a feed-forward neural network, which consid- ered as features the preceding nouns and determin- ers along with their part-of-speech tags. The win- ning systems of the 2016 task used deep neural networks: Luotolahti et al …

Tübingen system in VarDial 2017 shared task: Experiments with language identification and cross-lingual parsing
Ç Çöltekin, T Rama – Proceedings of the Fourth Workshop on NLP for …, 2017 – aclweb.org
… 2017). For this task, we experi- mented with two different families of models: lin- ear support vector machines (SVM), and (deep) neural network models … 2016). We experimented with both SVMs and (deep) neural network models …

Hierarchical Bayesian combination of plug-in maximum a posteriori decoders in deep neural networks-based speech recognition and speaker adaptation
Z Huang, SM Siniscalchi, CH Lee – Pattern Recognition Letters, 2017 – Elsevier
… Hierarchical Bayesian combination of plug-in maximum a posteriori decoders in deep neural networks-based speech recognition and speaker adaptation … Keywords. System combination. Bayesian learning. Sequential patterns. Deep neural networks. Automatic speech recognition …

CTC in the context of generalized full-sum HMM training
A Zeyer, E Beck, R Schlüter, H Ney – Proc. Interspeech, 2017 – researchgate.net
… Albert Zeyer, Eugen Beck, Ralf Schlüter, Hermann Ney Human Language Technology and Pattern Recognition, Computer Science Department, RWTH Aachen University, 52062 Aachen, Germany {zeyer, beck, schlueter, ney}@cs.rwth-aachen.de Abstract …

A batch noise contrastive estimation approach for training large vocabulary language models
Y Oualil, D Klakow – arXiv preprint arXiv:1708.05997, 2017 – arxiv.org
… r. Mohamed, N. Jaitly, A. Senior, V. Vanhoucke, P. Nguyen, TN Sainath, and B. Kings- bury, “Deep neural networks for acoustic … Training neural network language models on very large corpora,” in Proceedings of the Conference on Human Language Technology and Empirical …

Spectral-temporal receptive fields and MFCC balanced feature extraction for robust speaker recognition
JC Wang, CY Wang, YH Chin, YT Liu, ET Chen… – Multimedia Tools and …, 2017 – Springer
… Human Language Technology Department, Institute for Infocomm Research, A*STAR, SingaporeGoogle Scholar. 3. Anthony L, Kong AL, Bin M … L, Nicolas S, Luciana F, Mitchell M (2014) A novel scheme for speaker recognition using a phonetically-aware deep neural network …

Low-rank and sparse soft targets to learn better dnn acoustic models
P Dighe, A Asaei, H Bourlard – Acoustics, Speech and Signal …, 2017 – ieeexplore.ieee.org
… GE Dahl, A.-r. Mohamed, N. Jaitly, A. Senior, V. Vanhoucke, P. Nguyen, TN Sainath et al., “Deep neural networks for acoustic … and PC Woodland, “Tree-based state tying for high accuracy acoustic modelling,” in Proceedings of the workshop on Human Language Technology …

Discriminative information retrieval for question answering sentence selection
T Chen, B Van Durme – Proceedings of the 15th Conference of the …, 2017 – aclweb.org
… This research benefited from support by a Google Faculty Award, the JHU Human Language Technology Center of Excellence (HLTCOE), and DARPA DEFT … 2015. Learning to rank short text pairs with convolutional deep neural networks …

Learning profiles in duplicate question detection
C Saedi, J Rodrigues, J Silva… – Information Reuse and …, 2017 – ieeexplore.ieee.org
… from the Meta forum in StackExchange and AskUbuntu forum, with an 80%/20% train/test split.3 A deep neural network approach to … the effect of training corpus size on classifier performance for NLP,” in Proceedings of the 1st Human Language Technology (HLT) Conference …

Hybrid neural network alignment and lexicon model in direct hmm for statistical machine translation
W Wang, T Alkhouli, D Zhu, H Ney – … of the 55th Annual Meeting of the …, 2017 – aclweb.org
… Weiyue Wang, Tamer Alkhouli, Derui Zhu, Hermann Ney Human Language Technology and Pattern Recognition, Computer Science Department RWTH Aachen University, 52056 Aachen, Germany @i6.informatik.rwth-aachen.de Abstract …

Graph-based Neural Multi-Document Summarization
M Yasunaga, R Zhang, K Meelu, A Pareek… – arXiv preprint arXiv …, 2017 – arxiv.org
… redundancy. In our experiments on DUC 2004, we consider three types of sentence relation graphs and demonstrate the advantage of combining sentence relations in graphs with the repre- sentation power of deep neural networks …

Encoding syntactic knowledge in neural networks for sentiment classification
M Huang, Q Qian, X Zhu – ACM Transactions on Information Systems …, 2017 – dl.acm.org
Page 1. 26 Encoding Syntactic Knowledge in Neural Networks for Sentiment Classification MINLIE HUANG*, QIAO QIAN*, and XIAOYAN ZHU, State Key Laboratory of Intelligent Technology and Systems, National Laboratory …

Aroma: A recursive deep learning model for opinion mining in arabic as a low resource language
A Al-Sallab, R Baly, H Hajj, KB Shaban… – ACM Transactions on …, 2017 – dl.acm.org
Page 1. 25 AROMA: A Recursive Deep Learning Model for Opinion Mining in Arabic as a Low Resource Language AHMAD AL-SALLAB, Electronics and Communications Dpt., Faculty of Engineering, Cairo University RAMY …

Optimal hyperparameters for deep lstm-networks for sequence labeling tasks
N Reimers, I Gurevych – arXiv preprint arXiv:1707.06799, 2017 – arxiv.org
… However, it does not help to prioritize which parameter choices and extensions to implement in the first place. More practical recommendations for training deep neural network architectures and selecting hyperparameters is given by Bengio (2012) …

Identifying customer needs from user-generated content
A Timoshenko, JR Hauser – 2017 – papers.ssrn.com
Page 1. Electronic copy available at: https://ssrn.com/abstract=2985759 Identifying Customer Needs from User-Generated Content by Artem Timoshenko and John R. Hauser April 2017 Artem Timoshenko is a PhD student at …

Investigations on byte-level convolutional neural networks for language modeling in low resource speech recognition
K Irie, P Golik, R Schlüter, H Ney – Acoustics, Speech and …, 2017 – ieeexplore.ieee.org
… Human Language Technology and Pattern Recognition, Computer Science Department, RWTH Aachen University, Aachen, Germany {irie, golik, schlueter, ney}@cs.rwth … [3] Z. Tüske, P. Golik, R. Schlüter, and H. Ney, “Acous- tic Modeling with Deep Neural Networks Using Raw …

Researching mental health disorders in the era of social media: Systematic review
A Wongkoblap, MA Vadillo, V Curcin – Journal of medical Internet …, 2017 – ncbi.nlm.nih.gov

Dialect Recognition Based on Unsupervised Bottleneck Features
Q Zhang, JHL Hansen – … 2017, 18th Annual Conference of the …, 2017 – isca-speech.org
… [14] George E Dahl, Dong Yu, Li Deng, and Alex Acero, “Context- dependent pre-trained deep neural networks for large-vocabulary speech recognition,” IEEE Transactions on Audio, Speech, and Language Processing, vol … the workshop on Human Language Technology …

Navigation-orientated natural spoken language understanding for intelligent vehicle dialogue
Y Zheng, Y Liu, JHL Hansen – Intelligent Vehicles Symposium …, 2017 – ieeexplore.ieee.org
… Deep learning has been successfully applied to a number of human language technology areas including language modeling [23, 24], especially with the use of a natural deep framework of Recurrent Neural Networks (RNN) [26] … III. III. DEEP NEURAL NETWORKS …

Multitask Learning with Low-Level Auxiliary Tasks for Encoder-Decoder Based Speech Recognition
S Toshniwal, H Tang, L Lu, K Livescu – arXiv preprint arXiv:1704.01631, 2017 – arxiv.org
… [2] K. Vesel`y, A. Ghoshal, L. Burget, and D. Povey, “Sequence- discriminative training of deep neural networks.” in Interspeech … with neural networks,” in North American Chapter of the Association for Computational Linguistics on Human Language Technology (NAACL HLT …

A Deep Neural Network Approach To Parallel Sentence Extraction
F Grégoire, P Langlais – arXiv preprint arXiv:1709.09783, 2017 – arxiv.org
Page 1. A Deep Neural Network Approach To Parallel Sentence Extraction … In this paper, we propose a deep neural network approach to parallel sentence extraction that takes as input a pair of documents and outputs sentence pairs classified as transla- tions of each other …

JHU Kaldi system for Arabic MGB-3 ASR challenge using diarization, audio-transcript alignment and transfer learning
V Manohar, D Povey, S Khudanpur – … Speech Recognition and …, 2017 – danielpovey.com
… Center for Language and Speech Processing, Human Language Technology Center Of Excellence, Johns Hopkins University, Baltimore MD {vimal.manohar91,dpovey … in [8]. This suggests that even out-of-domain data is very critical for the training deep neural networks …

MorphoRuEval-2017: an evaluation track for the automatic morphological analysis methods for Russian
A Sorokin, T Shavrina, O Lyashevskaya, B Bocharov… – 2017 – dspace.spbu.ru
Page 1. Computational Linguistics and Intellectual Technologies: Proceedings of the International Conference “Dialogue 2017” Moscow, May 31—June 3, 2017 MorphoruEval-2017: an Evaluation track for thE autoMatic Morphological analysis MEthods for russian …

A Deep Learning Framework for Coreference Resolution Based on Convolutional Neural Network
JL Wu, WY Ma – Semantic Computing (ICSC), 2017 IEEE 11th …, 2017 – ieeexplore.ieee.org
… 1405–1415. [6] XF Xi, G. Zhou, F. Hu, and B. Fu, “A convolutional deep neural network for coreference resolution via modeling hierarchical features,” Proc … the Human Language Technology and North American Assoc. for Computational Linguistics (NAACL-HLT 2016), Jun …

Recognizing Text Entailment via Bidirectional LSTM Model with Inner-Attention
C Sun, Y Liu, C Jia, B Liu, L Lin – International Conference on Intelligent …, 2017 – Springer
… In: Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp … 1480–1489 (2016)Google Scholar. 9. Severyn, A., Moschitti, A.: Learning to rank short text pairs with convolutional deep neural networks …

A survey of multimodal sentiment analysis
M Soleymani, D Garcia, B Jou, B Schuller… – Image and Vision …, 2017 – Elsevier
Skip to main content …

Improving sentiment analysis with document-level semantic relationships from rhetoric discourse structures
J Märkle-Huß, S Feuerriegel… – … of the 50th …, 2017 – hl-128-171-57-22.library.manoa …
… For instance, [5] uses syntax and lexical information to obtain sentence- but not document- level semantic structures. Similarly, [6] studies the internal composition of sentences by considering the syntax tree structure with the help of recursive deep neural networks …

A Sentiment-and-Semantics-Based Approach for Emotion Detection in Textual Conversations
U Gupta, A Chatterjee, R Srikanth… – arXiv preprint arXiv …, 2017 – arxiv.org
… (c) Deep Learning approaches – Deep Neural networks have enjoyed … In Proceedings of the conference on human language technology and empirical methods in natural language processing, pages 579–586. ACL, 2005. [3] RC Balabantaray, M. Mohammad, and N. Sharma …

Position-aware Attention and Supervised Data Improve Slot Filling
Y Zhang, V Zhong, D Chen, G Angeli… – Proceedings of the 2017 …, 2017 – aclweb.org
Page 1. Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 35–45 Copenhagen, Denmark, September 7–11, 2017. cO2017 Association for Computational Linguistics Position …

Robust Coreference Resolution and Entity Linking on Dialogues: Character Identification on TV Show Transcripts
HY Chen, E Zhou, JD Choi – Proceedings of the 21st Conference on …, 2017 – aclweb.org
Page 1. Proceedings of the 21st Conference on Computational Natural Language Learning (CoNLL 2017), pages 216–225, Vancouver, Canada, August 3 – August 4, 2017. cO2017 Association for Computational Linguistics …

An extremely simple macroscale electronic skin realized by deep machine learning
KS Sohn, J Chung, MY Cho, S Timilsina, WB Park… – Scientific reports, 2017 – nature.com
… sensory device. A deep neural network (DNN) enabled us to process electrical resistance change induced by applied pressure and thereby to instantaneously evaluate the pressure level and the exact position under pressure. The …

De-identification of clinical notes via recurrent neural network and conditional random field
Z Liu, B Tang, X Wang, Q Chen – Journal of biomedical informatics, 2017 – Elsevier
Skip to main content …

Training deep networks to be spatially sensitive
N Kolkin, G Shakhnarovich… – arXiv preprint arXiv …, 2017 – openaccess.thecvf.com
… the number of pixels in the image. This computational cost poses a particular problem if the Fw ? metric were used as the training objective for deep neural networks (DNNs). Normally trained with stochas- tic gradient descent …

Composite Task-Completion Dialogue System via Hierarchical Deep Reinforcement Learning
B Peng, X Li, L Li, J Gao, A Celikyilmaz, S Lee… – arXiv preprint arXiv …, 2017 – arxiv.org
Page 1. Composite Task-Completion Dialogue System via Hierarchical Deep Reinforcement Learning Baolin Peng? Xiujun Li† Lihong Li† Jianfeng Gao† Asli Celikyilmaz† Sungjin Lee† Kam-Fai Wong? †Microsoft Research …

HCCL at SemEval-2017 Task 2: Combining Multilingual Word Embeddings and Transliteration Model for Semantic Similarity
J He, L Wu, X Zhao, Y Yan – … of the 11th International Workshop on …, 2017 – aclweb.org
… In Proceedings of the 2003 Conference of the North American Chapter of the Association for Computa- tional Linguistics on Human Language Technology- Volume 1 … Leverage financial news to predict stock price movements using word embeddings and deep neural networks …

Approximated and domain-adapted LSTM language models for first-pass decoding in speech recognition
M Singh, Y Oualil, D Klakow – Proceedings of the 18th …, 2017 – pdfs.semanticscholar.org
… Using other acoustic model training approaches such as deep neural networks is planned as a natural next step to improve our … BA Lund, and MA Przybocki, “1993 benchmark tests for the ARPA spoken language program,” in Human Language Technology, Pro- ceedings of a …

Acoustic feature learning via deep variational canonical correlation analysis
Q Tang, W Wang, K Livescu – arXiv preprint arXiv:1708.04673, 2017 – arxiv.org
… [6] L. Badino, C. Canevari, L. Fadiga, and G. Metta, “Integrating ar- ticulatory data in deep neural network-based acoustic modeling … Human Language Technology/Conference of the North American Chapter of the Association for Computational Linguis- tics (HLT/NAACL), 2015 …

The 2016 RWTH keyword search system for low-resource languages
P Golik, Z Tüske, K Irie, E Beck, R Schlüter… – … Conference on Speech …, 2017 – Springer
… Knill, KM, Gales, MJF, Rath, SP, Woodland, PC, Zhang, C., Zhang, SX: Investigation of multilingual deep neural networks for spoken term detection … author; Zoltán Tüske: 1. Kazuki Irie: 1. Eugen Beck: 1. Ralf Schlüter: 1. Hermann Ney: 1. 1.Human Language Technology and Pattern …

A continuously growing dataset of sentential paraphrases
W Lan, S Qiu, H He, W Xu – arXiv preprint arXiv:1708.00391, 2017 – arxiv.org
… This model achieved the best performance in the PIT- 2015 (Xu et al., 2014) dataset. DeepPairwiseWord He et al. (2016) developed a deep neural network model that focuses on im- portant pairwise word interactions across input sentences …

Chinese Medical Question Answer Matching Using End-to-End Character-Level Multi-Scale CNNs
S Zhang, X Zhang, H Wang, J Cheng, P Li, Z Ding – Applied Sciences, 2017 – mdpi.com
This paper focuses mainly on the problem of Chinese medical question answer matching, which is arguably more challenging than open-domain question answer matching in English due to the combination of its domain-restricted nature and the language-specific features of Chinese …

Automated detection of adverse drug reactions from social media posts with machine learning
I Alimova, E Tutubalina – International Conference on Analysis of Images …, 2017 – Springer
Adverse drug reactions can have serious consequences for patients. Social media is a source of information useful for detecting previously unknown side effects from a drug since users publish valuable.

Story Comprehension for Predicting What Happens Next
S Chaturvedi, H Peng, D Roth – Proceedings of the 2017 Conference on …, 2017 – aclweb.org
… 3.2 Baselines We use the following baselines in our experiments: DSSM: (Mostafazadeh et al., 2016) It trains two deep neural networks (Huang et al., 2013) to project the context and the ending-options into the same vector space …

Speech recognition for under-resourced languages: Data sharing in hidden Markov model systems
F de Wet, N Kleynhans, D Van Compernolle… – South African Journal …, 2017 – scielo.org.za
… A current popular trend for data sharing is to make use of deep neural networks (DNNs) for robust feature extraction, for which gains … The Afrikaans speech data that were used in this study were taken from the National Centre for Human Language Technology (NCHLT) speech …

Conclusion and Future Work
R Shah, R Zimmermann – … Analysis of User-Generated Multimedia Content, 2017 – Springer
… Due to advancements in computing power and deep neural networks (DNN), it is now feasible to quickly recognize a huge number of concepts in UGIs and UGVs … Advancements in deep neural networks help us in analyzing affective information from UGC …

A novel and robust approach for pro-drop language translation
L Wang, Z Tu, X Zhang, S Liu, H Li, A Way, Q Liu – Machine Translation, 2017 – Springer
… 4.2 DP generation. In light of the recent success of applying deep neural network technologies in natural language processing (Raymond and Riccardi 2007; Mesnil et al. 2013), we propose a neural network-based DP generator via the DP-inserted corpus …

Improving acoustic modeling using audio-visual speech
AH Abdelaziz – Multimedia and Expo (ICME), 2017 IEEE …, 2017 – ieeexplore.ieee.org
… E. Vincent, and D. Kolossa, “Uncertainty propagation through deep neural networks,” in Proc. Interspeech, 2015. [11] SJ Young, JJ Odell, and PC Woodland, “Tree-based state tying for high accuracy acoustic modelling,” in Proc. Workshop on Human Language Technology …

Semantic Sentiment Analysis of Twitter Data
P Nakov – arXiv preprint arXiv:1710.01492, 2017 – arxiv.org
… 8 Preslav Nakov Semi-supervised learning. We should note two things about the use of deep neural networks. First they can often do quite well without the need for explicit fea- ture modeling, as they can learn the relevant features in their hidden layers starting from the raw text …

Extended Variability Modeling and Unsupervised Adaptation for PLDA Speaker Recognition
A McCree, G Sell… – Proc. Interspeech …, 2017 – pdfs.semanticscholar.org
… Human Language Technology Center of Excellence Johns Hopkins University, Baltimore, MD, USA alan.mccree@jhu.edu, gsell@jhu.edu, dgromero@jhu … Discriminant Analysis (PLDA) [2, 3, 4]. While much recent work has focused on us- ing Deep Neural Networks to improve …

Interweaving Domain Knowledge and Unsupervised Learning for Psychiatric Stressor Extraction from Clinical Notes
OR Zhang, Y Zhang, J Xu, K Roberts, XY Zhang… – … Conference on Industrial …, 2017 – Springer
… 965 (2013)CrossRefGoogle Scholar. 4. Collobert, R., Weston, J.: A unified architecture for natural language processing: deep neural networks with multitask … In: Proceedings of the Conference on Human Language Technology and Empirical Methods in Language Processing, pp …

On the use of vector representation for improved accuracy and currency of Twitter POS Tagging
D Samuel – 2017 – dalspace.library.dal.ca
… Type I system having much lower computational complexity than its Type II counter- parts. Large vocabularies and Deep Neural Networks have made for computationally intensive (sometimes prohibitively so) systems, and this has fuelled advancement in …

End-to-End System for Bacteria Habitat Extraction
F Mehryary, K Hakala, S Kaewphan, J Björne… – BioNLP 2017, 2017 – aclweb.org
… habitat extraction. The pipeline utilizes deep neural networks, con- ditional random field classifiers and vector space models to solve the various subtasks and the code is freely available at https://github.com/ TurkuNLP/BHE. In …

Ensembles of deep lstm learners for activity recognition using wearables
Y Guan, T Plötz – Proceedings of the ACM on Interactive, Mobile …, 2017 – dl.acm.org
… 2015]. Most popular in the field of human activity recognition using wearables are two variants of deep learning methods: i) Deep Convolutional Neural Networks (CNNs); and ii) Recurrent Deep Neural Networks, such as Long Short Term Memory (LSTM) networks …

Learning multiple layers of knowledge representation for aspect based sentiment analysis
DH Pham, AC Le – Data & Knowledge Engineering, 2017 – Elsevier
… Glorot et al. [40] used a stacked denoising autoencoder to extract a semantic representation for each review. Socher et al. [41] proposed a family of recursive deep neural networks (RNN) to compute compositional vector representations for phrases. Le et al …

An investigation of deep-learning frameworks for speaker verification antispoofing
C Zhang, C Yu, JHL Hansen – IEEE Journal of Selected Topics …, 2017 – ieeexplore.ieee.org
… More generally, 40 dimensional filter- bank features with a deep neural network (DNN) as a back-end classifier was reported to be effective in spoofing detection [27] … All models are implemented with with the toolkit: Theano [43]. A. Deep Neural Networks …

Phonetic state relation graph regularized deep neural network for robust acoustic model
H Chung, YR Oh, SJ Lee… – Neural Networks (IJCNN) …, 2017 – ieeexplore.ieee.org
… [7] VS Tomar and RC Rose, “Manifold regularized deep neural networks.” in INTERSPEECH, 2014, pp … [14] SJ Young, JJ Odell, and PC Woodland, “Tree-based state tying for high accuracy acoustic modelling,” in Proceedings of the workshop on Human Language Technology …

Generating Fake but Realistic Headlines Using Deep Neural Networks
A Dandekar, RAM Zen, S Bressan – International Conference on Database …, 2017 – Springer
… In this paper, our aim is to use deep neural networks to generate the text and hence evaluate the quality of synthetically generated text against its topical coherence as well … In: Proceedings of the Second International Conference on Human Language Technology Research, pp …

Academic Activities Transaction Extraction Based on Deep Belief Network
X Wang, F Huang, W Wan, C Zhang – Advances in Multimedia, 2017 – hindawi.com
… elements. DBN is a deep neural network learning model and its hidden layer of neural network has excellent feature learning ability and can learn from the shallow features of the raw data to the deep level abstract features …

Named entity recognition for Amharic using deep learning
B Gambäck, UK Sikdar – IST-Africa Week Conference (IST …, 2017 – ieeexplore.ieee.org
… Recently, deep neural networks have been shown to effectively solve several language processing tasks such as part-of-speech tagging, sentiment analysis … In Proceedings of Conference on Human Language Technology for Development, pages 94–99, Alexandria, Egypt, May …

Philosophische Fakultät
D de Kok, R Ludwig – sfs.uni-tuebingen.de
… another. In this way we get a deep neural network with several hidden layers. The effect is, that every layer abstracts from the layer below … the North American Chapter of the Association for Computational Linguistics on Human Language Technology-Volume 1 (pp. 24-31) …

Headline Generation using Deep Neural Networks
D Vora – 2017 – scholarworks.sjsu.edu
… Spring 5-22-2017 Headline Generation using Deep Neural Networks Dhruven Vora San Jose State University … For more information, please contact scholarworks@sjsu.edu. Recommended Citation Vora, Dhruven, “Headline Generation using Deep Neural Networks” (2017) …

GradAscent at EmoInt-2017: Character and Word Level Recurrent Neural Network Models for Tweet Emotion Intensity Detection
E Lakomkin, C Bothe, S Wermter – Proceedings of the 8th Workshop on …, 2017 – aclweb.org
… Character-level deep neural networks recently showed outstanding results on text understanding tasks such as machine translation (Kalchbrenner … In HLT/EMNLP 2005, Human Language Technology Conference and Con- ference on Empirical Methods in Natural Lan- guage …

Prayas at EmoInt 2017: An Ensemble of Deep Neural Architectures for Emotion Intensity Prediction in Tweets
P Goel, D Kulshreshtha, P Jain, KK Shukla – Proceedings of the 8th …, 2017 – aclweb.org
… 2013. Cross-language knowledge transfer using multilingual deep neural network with shared hidden layers … In Proceedings of the con- ference on human language technology and empiri- cal methods in natural language processing …

EXPERIMENTS ON DIFFERENT RECURRENT NEURAL NETWORKS FOR ENGLISH-HINDI MACHINE TRANSLATION
R Agrawal, DM Sharma – airccj.org
… inspection of samples that there is a significant improvement in performance over rule-based and statistical approaches by using deep neural networks, thereby producing … In Proceedings of the second international conference on Human Language Technology Research …

Anaphora resolution with pointer networks
C Lee, S Jung, CE Park – Pattern Recognition Letters, 2017 – Elsevier
… Deep neural networks are constructed with multiple levels of hidden layers, and each layer uses a nonlinear activation function that transforms each representation at a lower level into a representation at a higher and slightly more abstract level …

Semantic Relation Classification for Concepts of Wordnet with Deep Learning
K Batsuren, A Chagnaa, A Ganbold, Z Munkhjargal – mmt.edu.mn
… IV. CONCLUSION In this article we proposed the deep neural network for Semantic Relation Classification for Concepts of WordNet … Proceedings of the Human Language Technology Conference of the North American Chapter of the ACL(NAACL), 1-4, 2006. [2] Luong, M. et al …

CCG Supertagging via Bidirectional LSTM-CRF Neural Architecture
R Kadari, Y Zhang, W Zhang, T Liu – Neurocomputing, 2017 – Elsevier
… In Recent years, deep neural networks have increasingly been powerful for various NLP-related tasks. A number of deep neural networks based approaches have addressed the problem of CCG supertagging such as Recurrent …

AI-Powered Social Bots
T Adams – arXiv preprint arXiv:1706.05143, 2017 – arxiv.org
… See [21] and [22] for interesting output from some trained generative networks. VI. MULTIMODAL PERSONAS & SCARY BOTS Deep neural networks are being applied in just about every human language technology to achieve state of the art for recognition …

Local Contexts Are Effective for Neural Aspect Extraction
J Yuan, Y Zhao, B Qin, T Liu – Chinese National Conference on Social …, 2017 – Springer
… In: Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp … Hinton, G., Deng, L., Yu, D., et al.: Deep neural networks for acoustic modeling in speech recognition: the shared views of four research groups …

Structure Regularized Bidirectional Recurrent Convolutional Neural Network for Relation Classification
J Wen – arXiv preprint arXiv:1711.02509, 2017 – arxiv.org
… Zeng et al. (2014) exploit a convolutional deep neural network to extract lexical and sentence level features. Wang et al … In Proceedings of Human Language Technology Conference and Confer- ence on Empirical Methods in Natural Language Process- ing, 724–731 …

A technical reading in statistical and neural machines translation (SMT & NMT)
Z El Maazouzi, BE El Mohajir… – … Technology (ICIT), 2017 …, 2017 – ieeexplore.ieee.org
… The progress of deep neural networks have sparked interest in applying them to natural language processing as well … & Ureš, L. (1994, March). The Candide system for machine translation. In Proceedings of the workshop on Human Language Technology (pp. 157-162) …

A Survey on Relation Extraction
M Cui, L Li, Z Wang, M You – China Conference on Knowledge Graph and …, 2017 – Springer
… In: Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp … Zeng, D., Liu, K., Lai, S., et al.: Relation classification via convolutional deep neural network. In: Proceedings of COLING, pp …

Introduction to the Special Issue on End-to-End Speech and Language Processing
B Ramabhadran, NF Chen, MP Harper… – IEEE Journal of …, 2017 – ieeexplore.ieee.org
… A joint speech dere- verberation for signal enhancement and a deep neural network based acoustic modeling architecture is presented … She is currently leading initiatives in deep learning and human language technology at the Institute for Infocomm Research, Agency for …

Based on Deep Belief Network
X Wang, F Huang, W Wan, C Zhang – downloads.hindawi.com
… elements. DBN is a deep neural network learning model and its hidden layer of neural network has excellent feature learning ability, and can learn from the shallow features of the raw data to the deep level abstract features …

Recognizing irregular entities in biomedical text via deep neural networks
F Li, M Zhang, B Tian, B Chen, G Fu, D Ji – Pattern Recognition Letters, 2017 – Elsevier
… Pattern Recognition Letters. Recognizing irregular entities in biomedical text via deep neural networks … As the techniques of deep learning develop in recent years, some researchers try to leverage deep neural networks to process biomedical NER tasks …

Social and Ethical Impact of Artificial Intelligence on Public-A Case Study of University Students
FF Quraishi, SA Wajid, P Dhiman – 2017 – ijsrset.com
… Papers in literature shows that deep neural networks are trained that produce basic radiological findings, results produced are … and Maxine Eskenazi, “Automatic Question Generation for Vocabulary Assessment,” Proceedings of Human Language Technology Conference and …

Study of speaker recognition system based on Feed Forward Deep Neural Networks exploring text-dependent mode
BJ Makrem, J Imen, O Kaïs – Information and Digital …, 2017 – ieeexplore.ieee.org
… Forward Deep Neural Networks exploring text … The RSR2015 database, designed to evaluate text- dependent speaker verification systems under different dura- tions and lexical constraints has been collected and released by the Human Language Technology (HLT) department …

Parallel Hierarchical Attention Networks with Shared Memory Reader for Multi-Stream Conversational Document Classification
N Sawada, R Masumura… – Proc. Interspeech …, 2017 – pdfs.semanticscholar.org
Page 1. Parallel Hierarchical Attention Networks with Shared Memory Reader for Multi-Stream Conversational Document Classification Naoki Sawada1,2, Ryo Masumura1, Hiromitsu Nishizaki3 1NTT Media Intelligence Laboratories …

Center-shared sliding ensemble of neural networks for syntax analysis of natural language
K Kim, Y Jin, SH Na, YK Kim – Expert Systems with Applications, 2017 – Elsevier
Skip to main content …

Learning Transferable Representation for Bilingual Relation Extraction via Convolutional Neural Networks
B Min, Z Jiang, M Freedman… – Proceedings of the Eighth …, 2017 – aclweb.org
… Recently, deep neural networks start to show promising re- sults in relation extraction … Besides relation extraction, (Huang et al., 2013) performs cross-language knowledge trans- fer with deep neural networks for speech recog- nition …

On the Use of Machine Translation-Based Approaches for Vietnamese Diacritic Restoration
TH Pham, XK Pham, P Le-Hong – arXiv preprint arXiv:1709.07104, 2017 – arxiv.org
… B. Neural-Based Machine Transslation In the last few years, deep neural network approaches have achieved state-of-the-art results … the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology-Volume 1 …

Sentence selection with neural networks using string kernels
M Masala, S Ruseti, T Rebedea – Procedia Computer Science, 2017 – Elsevier
… kernels alone or other unsupervised methods, but below the state of the art achieved with deep neural networks for answer … Radev, Using random walks for question-focused sentence retrieval, in: Proceedings of the conference on Human Language Technology and Empirical …

A Discourse-Level Named Entity Recognition and Relation Extraction Dataset for Chinese Literature Text
J Xu, J Wen, X Sun, Q Su – arXiv preprint arXiv:1711.07010, 2017 – arxiv.org
… Zeng et al. (2014) exploit a convolutional deep neural network to extract … In Proceedings of Human Language Technology Conference and Con- ference on Empirical Methods in Natural Language Pro- cessing, pages 724–731, Vancouver, British Columbia, Canada, October …

Improving Chemical-induced Disease Relation Extraction with Learned Features Based on Convolutional Neural Network
HQ Le, DC Can, TH Dang, MV Tran, QT Ha, N Collier – ieeexplore.ieee.org
… Relation classification via convolutional deep neural network. COLING 2014. [7] TH Nguyen, R Grishman (2015) … Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing; October 2005; Vancouver, Canada …

Big Data, Deep Learning–At the Edge of X-Ray Speaker Analysis
BW Schuller – International Conference on Speech and Computer, 2017 – Springer
… J., De Mazancourt, H.: “Where the data are coming from?” ethics, crowdsourcing and traceability for big data in human language technology … IEEE Access 2, 514–525 (2014)CrossRefGoogle Scholar. 7. Covington, P., Adams, J., Sargin, E.: Deep neural networks for youtube …

Improving Distantly Supervised Relation Extraction using Word and Entity Based Attention
S Jat, S Khandelwal, P Talukdar – akbc.ws
… A shortest path dependency kernel for relation extraction. In Proceedings of the conference on human language technology and empirical meth- ods in natural language processing, pages 724–731 … Relation classification via convolutional deep neural network …

Speech Emotion Recognition via Ensembling Neural Networks
D Luo, Y Zou, D Huang – pkusz.edu.cn
… China 2 Human Language Technology, Institute for Infocomm Research/A*STAR, Singapore 3IMSL Shenzhen Key Lab, PKU-HKUST Shenzhen Hong Kong Institution *E-mail: zouyx@pkusz.edu.cn, huang@i2r.a-star.edu.sg Abstract—Deep Neural Network (DNN) based …

A Multilayer Perceptron based Ensemble Technique for Fine-grained Financial Sentiment Analysis
MS Akhtar, A Kumar, D Ghosal, A Ekbal… – Proceedings of the …, 2017 – aclweb.org
… We develop three deep neural network architecture based models, viz … 2005. Recognizing Contextual Polarity in Phrase- level Sentiment Analysis. In Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Process- ing …

KIT-Conferences
MIAR Roedder – 2017 – isl.anthropomatik.kit.edu

Spanish Sign Language Recognition with Different Topology Hidden Markov Models
CD Mart?nez-Hinarejos… – Proc. Interspeech …, 2017 – pdfs.semanticscholar.org
… P. Hoffmann, “Recognizing contextual polarity in phrase-level sentiment analysis,” in Proceedings of the conference on human language technology and empirical … N. Jaitly, A. Senior, V. Vanhoucke, P. Nguyen, TN Sainath, and B. Kings- bury, “Deep neural networks for acoustic …

Recent Advances on Neural Headline Generation
SQ Shen, YK Lin, CC Tu, Y Zhao, ZY Liu… – Journal of Computer …, 2017 – Springer
Page 1. Ayana, Shen SQ, Lin YK et al. Recent advances on neural headline generation. JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY 32(4): 768–784 July 2017. DOI 10.1007/s11390-017-1758-3 Recent Advances on Neural Headline Generation …

Attention-Based Memory Network for Sentence-Level Question Answering
B Zhuang – Social Media Processing: 6th National Conference …, 2017 – books.google.com
… 2 Related Work Recent years, many deep neural networks have been proposed for the QA task [12, 13], which have accelerated the progress of QA system … In: Proceedings of the Second International Conference on Human Language Technology Research, pp. 399–404 …

Deep neural networks for identification of sentential relations
W Yin – 2017 – edoc.ub.uni-muenchen.de
Page 1. Deep Neural Networks for Identification of Sentential Relations Dissertation an der Fakultät f¨ur Mathematik, Informatik und Statistik … 1.1 Deep Neural Networks Neural networks are powerful statistical models within the machine learning area …

Recurrent Convolutional Neural Network for Relation Classification
J Wen – 2017 – shumingma.com
… Zeng et al. (2014) exploit a convolutional deep neural network to extract lexical and sentence level features. Wang et al … In Proceedings of the conference on human language technology and em- pirical methods in natural language processing, 724–731 …

FIRST STEP TOWARDS ENHANCING WORD EMBEDDINGS WITH PITCH ACCENT FEATURES FOR DNN-BASED SLOT FILLING ON RECOGNIZED TEXT
S Stehwien, NT Vu – essv2017.coli.uni-saarland.de
… the-art methods are evaluated on the benchmark dataset from the Air- line Travel Information Systems (ATIS) corpus [1] and yield around 95% F1-score using deep neural network (DNN) architectures … In In Proceedings of the ARPA Workshop on Human Language Technology …

Understanding Deep Learning Generalization by Maximum Entropy
G Zheng, J Sang, C Xu – arXiv preprint arXiv:1711.07758, 2017 – arxiv.org
… In Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology: Tutorials-Volume 5, pp … Opening the black box of deep neural networks via informa- tion …

The Kaldi OpenKWS System: Improving Low Resource Keyword Search
J Trmal, M Wiesner, V Peddinti, X Zhang… – Proc. Interspeech …, 2017 – isca-speech.org
… 1 Center for Language and Speech Processing, Johns Hopkins University, USA 2 Human Language Technology Center of Excellence, Johns Hopkins … 2.2. Deep Neural Network-based ASR Systems 2.2.1. Cross-Entropy trained BLSTM systems We train a hybrid hidden Markov …

Paraphrase Generation with Deep Reinforcement Learning
Z Li, X Jiang, L Shang, H Li – arXiv preprint arXiv:1711.00279, 2017 – arxiv.org
… The generator, built on the sequence-to- sequence learning framework, can generate paraphrases given a sentence. The teacher, modeled as a deep neural network, can decide whether the sentences are paraphrases of each other …

Deep fisher faces
H Hanselmann, S Yan, H Ney – British Machine …, 2017 – www-i6.informatik.rwth-aachen.de
… Human Language Technology and Pattern Recognition Group RWTH Aachen University Aachen, Germany … loss [21] and in [3]. In the latter the authors explore a way to directly translate the Linear Discriminant Analysis (LDA) [4] into a training criterion for deep neural networks …

Dialogue Act Recognition via CRF-Attentive Structured Network
Z Chen, R Yang, Z Zhao, D Cai, X He – arXiv preprint arXiv:1711.05568, 2017 – arxiv.org
… that in the scope of this work, we limit the number of hops to 1 to ease the computational cost. 2.3 Structured CRF-Attention Network Traditional attention networks have proven to be an effective ap- proach for embedding categorical inference within a deep neural network …

Shortcut Sequence Tagging
H Wu, J Zhang, C Zong – National CCF Conference on Natural Language …, 2017 – Springer
… 93.0. 6 Related Work. Skip connections have been widely used for training deep neural networks … Vaswani, A., Bisk, Y., Sagae, K., Musa, R.: Supertagging with LSTMs. In: Proceedings of the Human Language Technology Conference of the NAACL (2016)Google Scholar. 25 …

Adversarial generation of natural language
S Subramanian, S Rajeswar, F Dutil, C Pal… – Proceedings of the 2nd …, 2017 – aclweb.org
… A conditional version is also described that can generate sequences con- ditioned on sentence characteristics. 1 Introduction Deep neural networks have recently enjoyed some success at modeling natural language (Mikolov et al., 2010; Zaremba et al., 2014; Kim et al., 2015) …

Deep Semantic Role Labeling with Self-Attention
Z Tan, M Wang, J Xie, Y Chen, X Shi – arXiv preprint arXiv:1712.01586, 2017 – arxiv.org
… Y = X + Sub-Layer(X) (10) We then apply layer normalization (Ba, Kiros, and Hinton 2016) after the residual connection to stabilize the activa- tions of deep neural network. Position Encoding The attention mechanism itself cannot distinguish between different positions …

Voiceless Stop Consonant Modelling and Synthesis Framework Based on MISO Dynamic System
G Korvel, B Kostek – Archives of Acoustics, 2017 – degruyter.com
… T., Lu H., Kane J., Suni A., Vainio M., King S., Alku P. (2014), Voice source modelling using deep neural networks for statistical … Zió?ko B., Ga?ka J., Suresh M., Wilson R., Zió?ko M. (2009), Triphone statistics for Polish lan- guage, Human Language Technology: Challenges of the …

Accelerating deep neural network learning for speech recognition on a cluster of GPUs
G Cong, B Kingsbury, S Gosh, G Saon… – Proceedings of the …, 2017 – dl.acm.org
Page 1. Accelerating deep neural network learning for speech … ABSTRACT We train deep neural networks to solve the acoustic modeling prob- lem for large-vocabulary continuous speech recognition. We em- ploy distributed processing using a cluster of GPUs …

Challenges and Open Problems in Signal Processing: Panel Discussion Summary from ICASSP 2017 [Panel and Forum]
YC Eldar, AO Hero III, L Deng, J Fessler… – IEEE Signal …, 2017 – ieeexplore.ieee.org
… itera tive reconstruction algorithm and treat it as a sequence of processing steps akin to a deep neural network and then … paper and scientific awards for the contributions to AI, machine learning, multimedia signal processing, speech and human language technology, and their …

End-to-End Spoken Language Translation
M Guo, A Haque – stanford.edu
… Inspired by the Convolutional, Long Short-Term Memory Deep Neural Network (CLDNN) (Sainath et al., 2015) approach, we use a convolutional network to re- duce the temporal length of the input by using a learned convolutional filter bank …

Towards Document-Level Neural Machine Translation
L Miculicich Werlen – 2017 – infoscience.epfl.ch
… 2016). The results show only marginal improvement of pronoun translation respect to a baseline SMT system. The systems with the best performance used deep neural networks: Luotolahti et al. (2016) and Dabre et al. (2016 …

REDUNDANT CODING AND DECODING OF MESSAGES IN HUMAN SPEECH COMMUNICATION
H Hermansky – pdfs.semanticscholar.org
… Figure 2 Multiband deep neural network for estimating posterior probabilities of speech sounds, which exploits spectral and temporal redundancies in … supported by the DARPA RATS program, by the IARPA BABEL program, by the JHU Human Language Technology Center of …

Task-specific Word Identification from Short Texts Using a Convolutional Neural Network
S Yuan, X Wu, Y Xiang – arXiv preprint arXiv:1706.00884, 2017 – arxiv.org
… The rest of this paper is organized as follows. In section 2, we first briefly review of the related work on feature selection and sentiment word identifi- cation, along with research on deep neural networks for short text modeling … 2.2 Deep neural networks for short text modeling …

A neural network multi-task learning approach to biomedical named entity recognition
G Crichton, S Pyysalo, B Chiu… – BMC …, 2017 – bmcbioinformatics.biomedcentral …
… classification. Liu et al. [21] used multi-task deep neural networks to learn representations for information retrieval and semantic classification by jointly training a model for both tasks which has shared and private layers. Their …

Punctuaction Reconstruction For Auto-Matic Speech Transcription
T Š?AVNICKÝ – dspace.vutbr.cz
Page 1. BRNO UNIVERSITY OF TECHNOLOGY VYSOKÉ U?ENÍ TECHNICKÉ V BRN? FACULTY OF INFORMATION TECHNOLOGY FAKULTA INFORMA?NÍCH TECHNOLOGIÍ DEPARTMENT OF COMPUTER GRAPHICS AND MULTIMEDIA …

Incremental Graph-based Neural Dependency Parsing
X Zheng – Proceedings of the 2017 Conference on Empirical …, 2017 – aclweb.org
… Deep neural networks were used to replace the classifiers for predicting optimal tran- sitions in transition-based parers (Chen and Man- ning, 2014) or the scoring functions for ranking the subgraphs in graph-based rivals (Kiperwass- er and Goldberg, 2016a,b). There are …

Implicit opinion analysis: Extraction and polarity labelling
HH Huang, JJ Wang, HH Chen – Journal of the Association for …, 2017 – Wiley Online Library
… The resulting implicit opinions are more difficult to extract and label than explicit ones. In this paper, cutting-edge machine-learning approaches – deep neural network and word-embedding – are adopted for implicit opinion mining at the snippet and clause levels …

The limitations of data perturbation for ASR of learner data in under-resourced languages
J Badenhorst, F De Wet – … of South Africa and Robotics and …, 2017 – ieeexplore.ieee.org
… Febe de Wet Human Language Technology Research Group CSIR Meraka Institute Pretoria, South Africa febe.dewet@gmail.com … We present results for subspace Gaussian mixture models (SGMMs) and deep neural networks (DNNs) implemented using Kaldi …

Learning local and global contexts using a convolutional recurrent network model for relation classification in biomedical text
D Raj, S SAHU, A Anand – Proceedings of the 21st Conference on …, 2017 – aclweb.org
Page 1. Proceedings of the 21st Conference on Computational Natural Language Learning (CoNLL 2017), pages 311–321, Vancouver, Canada, August 3 – August 4, 2017. cO2017 Association for Computational Linguistics …

ENGLISH-HINDI USING RNN’S
R Agrawal, DM Sharma – pdfs.semanticscholar.org
… inspection of samples that there is a significant improvement in performance over rule-based and statistical approaches by using deep neural networks, thereby producing … In Proceedings of the second international conference on Human Language Technology Research …

A Content-Based Neural Reordering Model for Statistical Machine Translation
Y Pan, X Li, Y Yang, C Mi, R Dong, W Zeng – China Workshop on Machine …, 2017 – Springer
… In: Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics: HLT … 2013)Google Scholar. Liu, X., Gao, J., He, X., et al.: Representation learning using multi-task deep neural networks for semantic …

Information extraction with neural networks
JY Lee – 2017 – dspace.mit.edu
Page 1. Information Extraction with Neural Networks by Ji Young Lee Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science at the …

A Neural-Symbolic Approach to Natural Language Tasks
Q Huang, P Smolensky, X He, L Deng, D Wu – arXiv preprint arXiv …, 2017 – arxiv.org
… 1 arXiv:1710.11475v1 [cs.CL] 29 Oct 2017 Page 2. • in a Deep Neural Network (DNN) architecture … In Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology- Volume 1, pp. 173–180 …

Joint Emoji Classification and Embedding Learning
X Li, R Yan, M Zhang – Asia-Pacific Web (APWeb) and Web-Age …, 2017 – Springer
… We propose a matching approach using deep neural networks by utilizing emoji embeddings and observe that the performance of emoji … In: Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp …

A Customized Attention-Based Long Short-Term Memory Network for Distant Supervised Relation Extraction
D He, H Zhang, W Hao, R Zhang, K Cheng – Neural computation, 2017 – MIT Press
… Hence, much of the literature (Zeng, Liu, Lai, Zhou, & Zhao, 2014; Zeng et al., 2015; Xu et al., 2015; Lin, Shen, Liu, Luan, & Sun, 2016) adopts deep neural network models to avoid artificially designed features. In the distant …

Multi-sense based neural machine translation
Z Yang, W Chen, F Wang, B Xu – Neural Networks (IJCNN) …, 2017 – ieeexplore.ieee.org
… NMT system which is designed as an end-to-end paradigm, is recently introduced as a solution to machine translation [1], [2], [3], [4], [5], [6]. Unlike the traditional statistical machine translation [7], [8], the NMT model at- tempts to build a unified deep neural network to accomplish …

Coreference Resolution with Mention Representation Using a Convolutional Neural Network
SW Jeong, S Kim, H Kim – Advanced Science Letters, 2017 – ingentaconnect.com
… We expected that these errors will be able to be overcome because deep neural networks can map well words with same … 15. X. Luo, On coreference resolution performance metrics, Proceedings of the conference on Human Language Technology and Empirical Methods in Nat …

Finding Potential Business through Text Mining Techniques Based on Automotive Industry
JW Cai – 2017 – etd.lib.nsysu.edu.tw
… [22] R. Collobert and J. Weston, “A unified architecture for natural language processing: Deep neural networks with multitask … Recognizing contextual polarity in phrase-level sentiment analysis,” Proceedings of the Conference on Human Language Technology and Empirical …

A Survey of Distant Supervision Methods using PGMs
G Madan – arXiv preprint arXiv:1705.03751, 2017 – arxiv.org
… 2005. A shortest path dependency kernel for relation extrac- tion. In Proceedings of the conference on human language technology and empirical methods in nat- ural language processing … 2014. Relation classification via convolutional deep neural network. In COLING.

Agree to Disagree: Improving Disagreement Detection with Dual GRUs
S Hiray, V Duppada – arXiv preprint arXiv:1708.05582, 2017 – arxiv.org
… We have trained a deep neural network with fusion of lexical and word vector based features to achieve state of the art … of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology: companion volume …

Faster sequence training
A Zeyer, I Kulikov, R Schlüter… – Acoustics, Speech and …, 2017 – ieeexplore.ieee.org
… Human Language Technology and Pattern Recognition, Computer Science Department, RWTH Aachen University, 52062 Aachen, Germany {zeyer … Arnab Ghoshal, Lukás Burget, and Daniel Povey, “Sequence-discriminative training of deep neural networks.,” in INTERSPEECH …

Deep Stacking Networks for Low-Resource Chinese Word Segmentation with Transfer Learning
J Xu, X Sun, S Li, X Cai, B Wei – arXiv preprint arXiv:1711.01427, 2017 – arxiv.org
… Recently, deep neural networks have made significant improvements in Chinese word segmentation. These works aim at automatically extracting features. Col- lobert et al. [10] developed a general neural network architecture for sequence labelling tasks …

If You Can’t Beat Them Join Them: Handcrafted Features Complement Neural Nets for Non-Factoid Answer Reranking
D Bogdanova, J Foster, D Dzendzik, Q Liu – Proceedings of the 15th …, 2017 – aclweb.org
Page 1. Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages 121–131, Valencia, Spain, April 3-7, 2017. cO2017 Association for Computational Linguistics …

Sentiment Analysis in Today’s Trend: A Review
SKDHK Dhruba, J Kalita – … Journal Of Engineering And Computer Science, 2017 – ijecs.in
… Deep Neural network is the most recent advancement done in the neural networks which consist of multiple hidden layers … “Recognizing contextual polarity in phrase-level sentiment analysis.” In Proceedings of the conference on human language technology and empirical …

Parallel Text Acquisition and Comparison of SMT and NMT with Low Resources in English-Malay Machine Translation
YL Yeong, TP Tan, CK Lim, SK Mohammad – ona2017.khmernlp.org
… CNN are special feedforward neural networks that allows deep neural networks (DNN) to be built and modelled using local receptive fields, pooling, and weight sharing [4]. CNN is an … Proceedings of the Human Language Technology Conference, 127– 133, Edmonton, 2003 …

Attention-Based Memory Network for Sentence-Level Question Answering
P Liu, C Zhang, W Zhang, Z Zhan, B Zhuang – … National Conference on …, 2017 – Springer
… 2 Related Work. Recent years, many deep neural networks have been proposed for the QA task [12, 13], which have accelerated the progress of QA system … In: Proceedings of the Second International Conference on Human Language Technology Research, pp. 399–404 …

The I2R-NWPU Text-to-Speech System for Blizzard Challenge 2017
Y Lu, Z Zhang, C Yang, H Ming, X Zhu, Y Zhang… – festvox.org
… Human Language Technology Department, Institute for Infocomm Research, A*STAR, Singapore, 138632 †Shaanxi Provincial Key Laboratory of Speech and … Like our previous entry, we still adopt the general deep neural network (DNN) guided unit selection and waveform …

Spectral Graph Convolutional Networks for Part-of-Speech Tagging
S Demirel – 2017 – kola.opus.hbz-nrw.de
… DNN Deep Neural Network GCN Graph Convolutional Network GDA Gradient Descent Algorithm … There is no general definition for a Deep Neural Network (DNN) but it is safe to assume that a neural network is considered deep if it at least consists of two hidden layers or more …

Inferring Clinical Correlations from EEG Reports with Deep Neural Learning
TR Goodwin, SM Harabagiu – hlt.utdallas.edu
… afforded the neurologist by natural language. Consequently, to approach this problem, we present a deep neural network architecture which we refer to as the Deep Section Recovery Model (DSRM). Illustrated in Figure 1, the …

Reducing the computational complexity for whole word models
H Soltau, H Liao, H Sak – Automatic Speech Recognition and …, 2017 – ieeexplore.ieee.org
… ARPA Workshop on Human Language Technology, 1994 … Interspeech, 2014. [25] H. Liao, E. McDermott, and A. Senior, “Large scale deep neural network acoustic modeling with semi-supervised training data for youtube video transcription,” in Proc. ASRU, 2013 …

Parallel Neural Network Features for Improved Tandem Acoustic Modeling
Z Tüske, W Michel, R Schlüter… – Proc …, 2017 – www-i6.informatik.rwth-aachen.de
… Human Language Technology and Pattern Recognition, Computer Science Department, RWTH Aachen University, 52056 Aachen, Germany {tuske,michel,schlueter,ney}@cs.rwth-aachen … [38] TN Sainath et al., “Low-rank matrix factorization for deep neural network training with …

Singleton Detection for Coreference Resolution via Multi-window and Multi-filter CNN
K Li, H Huang, Y Guo, P Jian – China Workshop on Machine Translation, 2017 – Springer
… Zeng, D., Liu, K., Lai, S., Zhou, G., Zhao, J.: Relation classification via convolutional deep neural network. In: COLING (2014)Google Scholar. 16 … In: Proceedings of the Human Language Technology Conference of the NAACL, Companion Volume: Short Papers, pp. 57–60 …

Segment convolutional neural networks (Seg-CNNs) for classifying relations in clinical notes
Y Luo, Y Cheng, Ö Uzuner, P Szolovits… – Journal of the …, 2017 – academic.oup.com
Abstract. We propose Segment Convolutional Neural Networks (Seg-CNNs) for classifying relations from clinical notes. Seg-CNNs use only word-embedding features.

Biasing Attention-Based Recurrent Neural Networks Using External Alignment Information
T Alkhouli, H Ney – Proceedings of the Second Conference on Machine …, 2017 – aclweb.org
… Tamer Alkhouli and Hermann Ney Human Language Technology and Pattern Recognition Group Computer Science Department RWTH Aachen University D-52056 Aachen, Germany @i6.informatik.rwth-aachen.de Abstract …

Entity linking for tweets
P Basile, A Caputo – Encyclopedia with Semantic Computing and …, 2017 – World Scientific
Page 1. Entity linking for tweets Pierpaolo Basile*,‡ and Annalina Caputo †,§ *Department of Computer Science, University of Bari Aldo Moro Via E. Orabona 4, Bari, 70125, Italy † ADAPT Centre, Trinity College Dublin, Dublin …

Neural Cross-Lingual Entity Linking
A Sil, G Kundu, R Florian, W Hamza – arXiv preprint arXiv:1712.01813, 2017 – arxiv.org
Page 1. Neural Cross-Lingual Entity Linking Avirup Sil and Gourab Kundu and Radu Florian and Wael Hamza IBM Research AI 1101 Kitchawan Road Yorktown Heights, NY 10598 {avi, gkundu, raduf, whamza}@us.ibm.com Abstract …

It sounds like you have a cold! Testing voice features for the Interspeech 2017 Computational Paralinguistics Cold Challenge
M Huckvale, A Beke – Proc. Interspeech 2017, 2017 – isca-speech.org
… Neural network classifier The Microsoft Cognitive Toolkit CNTK [31] was used to train and test deep neural network classifiers … Prosodic Features of Parkinson’s disease Speech.“ In STIL-IX Brazilian Symposium in Information and Human Language Technology, 2nd Brazilian …

The role of linguistic and prosodic cues on the prediction of self-reported satisfaction in contact centre phone calls
J Luque, C Segura, A Sánchez… – Proc. Interspeech …, 2017 – pdfs.semanticscholar.org
… Index Terms: Speech recognition, deep neural networks, cus- tomer satisfaction index (CSI), natural language processing 1. Introduction … of the North American Chapter of the Association for Computational Linguistics on Human Language Technology: Companion Volume …

Distributional Semantics and Neural Network based Improvements to Dependency Parsing
S Kanneganti – 2017 – web2py.iiit.ac.in
Page 1. Distributional Semantics and Neural Network based Improvements to Dependency Parsing Thesis submitted in partial fulfillment of the requirements for the degree of MS by Research in Computer Science & Engineering by Silpa Kanneganti 200602012 …

Exploring global sentence representation for graph-based dependency parsing using BLSTM-SCNN
N Si, H Wang, Y Shan – Pattern Recognition Letters, 2017 – Elsevier
… In this paper, we propose an effective deep neural network model for graph-based dependency parsing … Recently, deep neural networks such as CNN [13,14], LSTM [15,16], and attention based memory network [17], have shown good performance on dependency parsing …

Beta-Masking MMSE Speech Enhancement for Speech Recognition
C You, B Ma – 2017 – oar.a-star.edu.sg
… Chang Huai You, Bin Ma Human Language Technology, Institute for Infocomm Research (I2R) A?STAR, 1 Fusionopolis Way, Singapore Emails … [16] K. Vesely, A. Ghoshal, L. Burget, and D. Povey, “Sequence- discriminative training of deep neural networks,” Interspeech, 2013 …

Improving Black-box Speech Recognition using Semantic Parsing
R Corona, J Thomason, R Mooney – Proceedings of the Eighth …, 2017 – aclweb.org
… Collect- ing sufficient data to train ASR systems using cur- rent state of the art methods, such as deep neural networks (Graves and Jaitly, 2014; Xiong et al., 2016), is difficult … In Proceedings of the workshop on Human Language Technology, pages 272–277 …

TEDxSK and JumpSK: A New Slovak Speech Recognition Dedicated Corpus
J Staš, D Hládek, P Viszlay, T Koctúr – Journal of Linguistics …, 2017 – degruyter.com
… tract length normalization (VTLN) [9], speaker adaptive training (SAT) [10], or statistical modeling with deep neural networks (DNN) [11] … In Vertulani, Z., Uszkoreit, H., and Kubis, M., editors, Human Language Technology: Challenges for Computer Science and Linguistics, LNAI …

Context incorporation using context—aware language features
A Vlachostergiou, G Marandianos… – … (EUSIPCO), 2017 25th …, 2017 – ieeexplore.ieee.org
… context when processing sentences or paragraphs in isolation, underling the need of context incorporation to advance the human language technology systems wrt … Finally, we intend to compare our context – aware incorporation method with Deep Neural network approaches …

Dissertations in Forestry and Natural Sciences
A SIZOV – epublications.uef.fi
… Staff Douglas A. Reynolds, Ph.D. Lincoln Laboratory Massachusetts Institute of Technology Human Language Technology Group 244 … BTAS Biometrics: theory, applications, and systems CLDNN Convolutional, long short-term memory and deep neural network CQCC Constant …

Co-Morbidity Exploration on Wearables Activity Data Using Unsupervised Pre-training and Multi-Task Learning
K Aggarwal, S Joty, LF Luque, J Srivastava – arXiv preprint arXiv …, 2017 – arxiv.org
… 4.2 Supervised Multi-task Learning In recent years, deep neural networks (DNNs) have shown impres- sive performance gains in a wide spectrum of machine learning problems such as image recognition, language translation, speech recognition, natural language parsing …

NTCD-TIMIT: A New Database and Baseline for Noise-robust Audio-visual Speech Recognition
AH Abdelaziz – Proc. Interspeech 2017, 2017 – pdfs.semanticscholar.org
… Models The speaker-independent acoustic deep neural network/hidden Markov model (DNN/HMM) hybrid models have been trained as follows: Following the convention, mono-phone Gaussian mixture model/hidden Markov model (GMM/HMM) models have been firstly trained …

Attentive Language Models
G Salton, R Ross, J Kelleher – Proceedings of the Eighth International …, 2017 – aclweb.org
… Another interpretation of the smoothing effect is that it “reinforces” the signal in a similar fashion to residual connections in other RNNs and Deep Neural Networks architectures … In Proceedings of the Workshop on Human Language Technology, pages 114–119 …

Towards efficient Neural Machine Translation for Indian Languages
R Agrawal – 2017 – pdfs.semanticscholar.org
… The thesis proposes the use of Neural Machine Translation techniques for the task of Indian Language MT. Neural Machine Translation (NMT) is a novel approach to MT which utilizes deep neural networks to generate end-to-end translation …

Combining local and global features in supervised word sense disambiguation
X Lei, Y Cai, Q Li, H Xie, H Leung, FL Wang – International Conference on …, 2017 – Springer
… 2014)Google Scholar. 8. Collobert, R., Weston, J.: A unified architecture for natural language processing: deep neural networks with multitask … In: Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp …

Multi-task Domain Adaptation for Sequence Tagging
N Peng, M Dredze – Proceedings of the 2nd Workshop on …, 2017 – aclweb.org
… Multi-task Domain Adaptation for Sequence Tagging Nanyun Peng and Mark Dredze Human Language Technology Center of Excellence Center for Language and Speech Processing Johns Hopkins University, Baltimore, MD, 21218 npeng1@jhu.edu, mdredze@cs.jhu.edu …

End-to-End Architectures for Speech Recognition
Y Miao, F Metze – New Era for Robust Speech Recognition, 2017 – Springer
… Multiple training stages. The pipeline is initialized by constructing a hidden Markov model/Gaussian mixture model (HMM/GMM). It generates frame-level alignments of the training data, which are used as targets for the deep-neural-network (DNN) training …

Challenges in Building Highly Interactive Dialogue Systems
AI Magazine – AI Magazine, 2017 – cs.utep.edu
… Convolutional and deep neural networks have proven very useful for analogous problems in vision and other tasks, where they are able to model both low-level features and higher-level … In Cole, R., ed., Survey of the State of the Art in Human Language Technology, 204–210 …

Factored front-end CMLLR for joint speaker and environment normalization under DNN-HMM
SP Rath – International Journal of Speech Technology, 2017 – Springer
… Replacing Gaussian mixture model hidden Markov model (GMM–HMM), deep neural network HMM (DNN-HMM) has become the state of the art for acoustic modeling … Feature engineering in context-dependent Deep Neural Networks for conversational speech transcription …

Dependency Parsing with Dilated Iterated Graph CNNs
E Strubell, A McCallum – arXiv preprint arXiv:1705.00403, 2017 – arxiv.org
… vastly accelerating and parallelizing the core numeric operations for performing inference and computing gradients in neural networks, recent de- velopments in GPU hardware have facilitated the emergence of deep neural networks as state … Human Language Technology Conf …

Use of Knowledge Graph in Rescoring the N-Best List in Automatic Speech Recognition
AJ Kumar, C Morales, ME Vidal, C Schmidt… – arXiv preprint arXiv …, 2017 – arxiv.org
… with cleaned automatic transcripts 1. Mel frequency cepstral coefficients are used as features to train a deep neural network based acoustic … knowledge sources to reorder n-best speech hypoth- esis lists.” In Proceedings of the workshop on Human Language Technology, pp …

The Regional Style Classification of Chinese Folk Songs Based on GMM-CRF Model
J Li, J Ding, X Yang – Proceedings of the 9th International Conference on …, 2017 – dl.acm.org
… audio and regional tags. Deep Neural Network (DNN) was utilized as a classifier. Among them, the best result (72.9%) of folk songs regional style classification was based on correlation analysis of CCA. 2.2 Existing Problems …

A semantic annotation framework for scientific publications
Y Jung – Quality & Quantity, 2017 – Springer
… To achieve additional accuracy in named entity recognition task, we can think of a deep neural network (Dos Santos and Guimarães 2014 … of the 2003 conference of the North American chapter of the association for computational linguistics on human language technology, vol …

Statistical translation of English texts to API code templates
AT Nguyen, PC Rigby, T Van Nguyen… – … Companion (ICSE-C …, 2017 – ieeexplore.ieee.org
… based statistical machine translation (SMT) [9], probabilistic CFG [10], AST-based translation [11], and deep neural network [3]. Compared to … Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology – Volume 1 …

Impact of data augmentations when training the Inception model for image classification
M Barai, A Heikkinen – 2017 – diva-portal.org
Page 1. IN DEGREE PROJECT COMPUTER ENGINEERING, FIRST CYCLE, 15 CREDITS , STOCKHOLM SWEDEN 2017 Impact of data augmentations when training the Inception model for image classification ANTHONY HEIKKINEN MILAD BARAI …

Global Relation Embedding for Relation Extraction
Y Su, H Liu, S Yavuz, I Gur, H Sun, X Yan – arXiv preprint arXiv …, 2017 – arxiv.org
… Abstract Recent studies have shown that embed- ding textual relations using deep neural networks greatly helps relation extraction. However, many existing studies rely on supervised learning; their performance is dramatically limited by the availability of training data …

Improving Feature-Rich Transition-Based Constituent Parsing Using Recurrent Neural Networks
A TAMURA, L LIU, T ZHAO, E SUMITA – IEICE TRANSACTIONS on …, 2017 – search.ieice.org
Page 1. IEICE TRANS. INF. & SYST., VOL.E100–D, NO.9 SEPTEMBER 2017 2205 PAPER Improving Feature-Rich Transition-Based Constituent Parsing Using Recurrent Neural Networks Chunpeng MA †?a) , Akihiro TAMURA …

Overview of TAC-KBP2017 13 Languages Entity Discovery and Linking
H Ji, X Pan, B Zhang, J Nothman, J Mayfield… – nlp.cs.rpi.edu
… SAFT ISI USC Information Sciences Institute STANFORD Stanford University TinkerBell RPI, UIUC, Stanford, Columbia, Cornell, JHU, UPenn hltcoe Human Language Technology Center of Excellence newbie mr Machine Reading Co 2nd Evaluation Window 2089Pacific …

Natural Language Processing, Moving from Rules to Data
AH Dediu, JM Matos, C Martín-Vide – International Conference on Theory …, 2017 – Springer
… Microsoft Cortana, Amazon Echo, etc., marked a significant progress after 2012, mainly due to the recent advances in deep learning technologies [19] — especially deep neural networks (DNNs) … Cole, R. (ed.): Survey of the State of the Art in Human Language Technology …

Challenges in data-to-document generation
S Wiseman, SM Shieber, AM Rush – arXiv preprint arXiv:1707.08052, 2017 – arxiv.org
Page 1. arXiv:1707.08052v1 [cs.CL] 25 Jul 2017 Challenges in Data-to-Document Generation Sam Wiseman and Stuart M. Shieber and Alexander M. Rush School of Engineering and Applied Sciences Harvard University Cambridge …

Relating dynamic brain states to dynamic machine states: Human and machine solutions to the speech recognition problem
C Wingfield, L Su, X Liu, C Zhang… – PLoS computational …, 2017 – journals.plos.org
Author summary The ability to understand spoken language is a defining human capacity. But despite decades of research, there is still no well-specified account of how sound entering the ear is neurally interpreted as a sequence of meaningful words. At the same time, modern …

Word and Relation Embedding for Sentence Representation
T Rath – 2017 – search.proquest.com
… modelling. Below is a brief introduction about these topics.2.1 Articial Neural Networks. Neuron. A neuron is the basic computational unit in deep neural network architectures, which is loosely inspired by a biological neuron. It …

Prague at EPE 2017: The UDPipe System
M Straka, J Straková, J Hajic – EPE 2017, 2017 – svn.nlpl.eu
… Finally, we conclude in Section 5. 2 Related Work Deep neural networks have achieved remarkable results in many areas of machine learning … In Proceedings of the Second International Conference on Human Language Technology Research …

Relation extraction via one-shot dependency parsing on intersentential, higher-order, and nested relations
GG SAHIN, E EMEKLIGIL, S ARSLAN, O AGIN… – online.journals.tubitak.gov.tr
… [11] Zeng D, Liu K, Lai S, Zhou G, Zhao J. Relation classification via convolutional deep neural network … In: HLT ’05 Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing; 6-8 October 2005; Vancouver, BC …

Deep Reinforcement Learning in Natural Language Scenarios
J He – 2017 – digital.lib.washington.edu
… We address the first problem by introducing a bi-directional LSTM deep neural network architectures to account for the potential redundancy and/or temporal … It builds on recent work in deep reinforcement learning, specifically using deep neural networks as function approxima …

SSEL-ADE: A semi-supervised ensemble learning framework for extracting adverse drug events from social media
J Liu, S Zhao, G Wang – Artificial intelligence in medicine, 2017 – Elsevier
… Korkontzelos [14], DailyStrength& Twitter, Not applied, Feature- based, Feature- based, Supervised, F-measure: 80.14% (DS) and 69.16% (TW). Huynh et al. [30], Twitter, Deep Neural Networks, Not applied, Not applied, Supervised, F-measure: 51%; AUC: 88%. Lee et al …

Overview of the 2017 spoken CALL shared task
C Baur, C Chua, J Gerlach, E Rayner, M Russel, H Strik… – 2017 – archive-ouverte.unige.ch
… There are now many episodes in the history of human language technology showing that the introduction of a shared task1 has had a … Three baseline deep neural network (DNN)-hidden Markov model (HMM) hybrid systems were trained on the training part of WSJCAM0 [14 …

Comparing Fusion Models for DNN-based Audio-visual Continuous Speech Recognition
AH Abdelaziz – IEEE/ACM Transactions on Audio, Speech, and …, 2017 – ieeexplore.ieee.org
… However, very few studies can be found in the literature that compare different fusion models for AV-ASR. Even less research work compares audio-visual fusion models for large vocabulary continuous speech recognition (LVCSR) models using deep neural networks (DNNs) …

Movie description
A Rohrbach, A Torabi, M Rohrbach, N Tandon… – International Journal of …, 2017 – Springer
Audio description (AD) provides linguistic descriptions of movies and allows visually impaired people to follow a movie along with their peers. Such descriptions are by design mainly visual and thus n.

Improved Interpretability and Generalisation for Deep Learning
M Graziani – pdfs.semanticscholar.org
… 2 2 Deep Learning 3 2.1 Feedforward Deep Neural Networks . . . . . 3 … investigation within a broader context. 2.1 Feedforward Deep Neural Networks Feedforward Neural Networks (NNs) attempt to approximate a function f, which …

PELESent: Cross-domain polarity classification using distant supervision
EA Corrêa Jr, VQ Marinho, LB Santos… – arXiv preprint arXiv …, 2017 – arxiv.org
… Distant Supervision is a good alternative to obtain these datasets for the training/pre-training of deep neural networks [16], [18], [19] … of a portuguese opinion lexicon from multiple resources,” in 8th Brazilian Symposium in Information and Human Language Technology, 2011, pp …

Complex Structure Leads to Overfitting: A Structure Regularization Decoding Method for Natural Language Processing
X Sun, W Sun, S Ma, X Ren, Y Zhang, W Li… – arXiv preprint arXiv …, 2017 – arxiv.org
… Preprint submitted to Artificial Intelligence November 29, 2017 arXiv:1711.10331v1 [cs.LG] 25 Nov 2017 Page 2. the representative models are conditional random fields (CRFs), deep neural networks, and structured perceptron models …

Automatic Neural Question Generation using Community-based Question Answering Systems
T Baghaee – 2017 – uleth.ca
… users’ reviews. We first present a general understanding of the deep learning approach. We then go through the intuition behind deep neural networks and later expand this idea by describing more advanced models … 7 Page 17. 2.3. DEEP NEURAL NETWORK process …

Handling OOVWords in Mandarin Spoken Term Detection with an Hierarchical n-Gram Language Model
X Wang, P Zhang, X Na, J Pan, Y Yan – Chinese Journal of Electronics, 2017 – IET
… Although Deep neural networks (DNNs) make great contribution to LVCSR, the search for Out-of-vocabulary (OOV) words remains a … based state tying for high accuracy acoustic modelling”, Proceedings of the Work- shop on Human Language Technology, Plainsboro, New …

Coreference resolution system not only for Czech
M Novák – 2017 – pdfs.semanticscholar.org
… anaphoricity detection and antecedent selection, and specialized models [11]. A recent tsunami of deep neural network appears to be a small wave in the field of research on coreference. Neural Stanford system [8] set a new state …

Verbalization of Service Robot Experience as Explanations in Language Including Vision-Based Learned Elements
SPP Selvaraj – 2017 – pdfs.semanticscholar.org
… We use a deterministic approach to automatically find which floor the robot has entered without a human’s help. We demonstrate automatic annotation by using a deep neural network (DNN) to find which floor the CoBot is in after reaching a new floor via an elevator …

Robust Conditional Probabilities
Y Wald, A Globerson – Advances in Neural Information Processing …, 2017 – papers.nips.cc
… Entropy Regularizer: Consider training a deep neural network where the last layer has n neurons z1,…,zn connected to a softmax layer of size … In Proceedings of the main conference on human language technology conference of the North American Chapter of the Association of …

Automatic generation of named entity taggers leveraging parallel corpora
YL Chung – 2017 – addi.ehu.es
Page 1. Automatic Generation of Named Entity Taggers Leveraging Parallel Corpora Author: Yi-Ling Chung Advisors: Rodrigo Agerri and German Rigau Master’s Thesis Master in Language Analysis and Processing University …

Comparison study on critical components in composition model for phrase representation
S Wang, C Zong – ACM Transactions on Asian and Low-Resource …, 2017 – dl.acm.org
Page 1. 16 Comparison Study on Critical Components in Composition Model for Phrase Representation SHAONAN WANG, National Laboratory of Pattern Recognition, Institute of Automation, University of Chinese Academy …

Detecting annotation noise in automatically labelled data
I Rehbein, J Ruppenhofer – Proceedings of the 55th Annual Meeting of …, 2017 – aclweb.org
Page 1. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 1160–1170 Vancouver, Canada, July 30 – August 4, 2017. c 2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1107 …

Using Natural Language Processing Techniques for Stock Return Predictions
ML Chew, S Puri, A Sood, A Wearne – 2017 – papers.ssrn.com
Page 1. Electronic copy available at: https://ssrn.com/abstract=2940564 BlackRock Applied Finance Project: Using Natural Language Processing techniques for Stock Return Predictions Ming Li Chew, Sahil Puri, Arsh Sood, Adam Wearne March 27, 2017 1 Page 2 …

Multi-Modal and Deep Learning for Robust Speech Recognition
X Feng – 2017 – groups.csail.mit.edu
… Third, we developed a method to incorporate heterogeneous multi-modal data with a deep neural network (DNN) based acoustic model … Over the past few years the success of deep neural networks (DNNs) has further boosted the recognition performance, and has …

AraVec: A set of Arabic Word Embedding Models for use in Arabic NLP
AB Soliman, K Eissa, SR El-Beltagy – Procedia Computer Science, 2017 – Elsevier
… [12] R. Collobert and J. Weston, “A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask … from Text: Machine Learning for Text-based Emotion Prediction,” in Proceedings of the Conference on Human Language Technology and Empirical …

Multilingual sentiment analysis: from formal to informal and scarce resource languages
SL Lo, E Cambria, R Chiong, D Cornforth – Artificial Intelligence Review, 2017 – Springer
The ability to analyse online user-generated content related to sentiments (eg, thoughts and opinions) on products or policies has become a de-facto skillset for many companies and organisations. Be.

Tag Recommendation and Ranking
R Shah, R Zimmermann – … Analysis of User-Generated Multimedia Content, 2017 – Springer
… proposed [30, 101]. However, they have limited performance because classes (tags) used in training deep neural networks to predict tags for a UGI are restricted and defined by a few researchers and not by actual users. Thus, it …

VerbNet/OntoNotes-Based Sense Annotation
M Green, O Hargraves, C Bonial, J Chen… – Handbook of Linguistic …, 2017 – Springer
… 378–388. IEEE (2007)Google Scholar. 7. Collobert, R., Weston, J.: A unified architecture for natural language processing: deep neural networks with multitask learning … In: Proceedings of the Human Language Technology Workshop, Princeton (1994)Google Scholar. 35 …

Find the errors, get the better: Enhancing machine translation via word confidence estimation
NQ Luong, L Besacier, B Lecouteux – Natural Language Engineering, 2017 – cambridge.org
… Camargo-de-Souza et al. (2014). In WMT 2015, deep neural network is employed by Kreutzer, Schamoni and Riezler (2015) to learn continuous feature representations from bilingual contexts. Their network is firstly trained …

Inducing a Semantically Annotated Lexicon via Deep Variational Autoencoders and EM-Based Clustering
M Bleeker, T Scheepers, D Zomerdijk – thijs.ai
… clustered in the same class. Deep neural networks have taken the field of nat- ural language processing by storm and yielded large improvements for various language related tasks in the recent past. In section 5 we introduce …

Stance and sentiment in tweets
SM Mohammad, P Sobhani, S Kiritchenko – ACM Transactions on …, 2017 – dl.acm.org
Page 1. 26 Stance and Sentiment in Tweets SAIF M. MOHAMMAD, National Research Council Canada PARINAZ SOBHANI, University of Ottawa SVETLANA KIRITCHENKO, National Research Council Canada We can often …

Analysis of Images, Social Networks and Texts: 6th International Conference, AIST 2017, Moscow, Russia, July 27–29, 2017, Revised Selected Papers
WMP van der Aalst, DI Ignatov, M Khachay… – 2017 – books.google.com
… The keynote talk was presented by Andrzej Cichocki on “Bridge Between Tensor Networks and Deep Neural Networks: From Fundamentals to Real Applications.” The invited talks were: – Stanley Wasserman (Indiana University, USA), “Sensitivity Analysis of p* and SAOM: The …

SVitchboard-II and FiSVer-I: Crafting high quality and low complexity conversational english speech corpora using submodular function optimization
Y Liu, R Iyer, K Kirchhoff, J Bilmes – Computer Speech & Language, 2017 – Elsevier
… various submodular functions. • We provided baseline word recognition results on all of the resultant speech corpora for both Gaussian mixture model (GMM) and deep neural network (DNN)-based systems. • We had released …

Multitask feature learning for low-resource query-by-example spoken term detection
H Chen, CC Leung, L Xie, B Ma… – IEEE Journal of Selected …, 2017 – ieeexplore.ieee.org
… language. We extract low-dimensional features from a bottle-neck layer of a multitask deep neural network, which is jointly trained with speech data from the low-resource target language and resource-rich nontarget language …

Social Psychological Research Methods: Social Psychological Measurement
MS Chan, A Morales, M Farhadloo, RJ Palmer… – researchgate.net
Page 1. Social Psychological Research Methods: Social Psychological Measurement Harvesting and Harnessing Social Media Data for Psychological Research Man-pui Sally Chan Alex Morales Mohsen Farhadloo Ryan Joseph Palmer Dolores Albarracín In Blanton, H. et al …

Transfer Linear Subspace Learning for Cross-corpus Speech Emotion Recognition
P Song – IEEE Transactions on Affective Computing, 2017 – ieeexplore.ieee.org
… proposed in the literature, eg, sup- port vector machine (SVM), neural network (NN), hidden Markov model (HMM), Gaussian mixture model (GMM), regression algorithms, multilayer perception, decision trees, extreme learning machine (ELM), deep neural network (DNN) [1] …

How deep learning can help emotion recognition
PR Khorrami – 2017 – ideals.illinois.edu
… enced a resurgence mainly due to the success of deep neural networks. In this dissertation, we highlight how deep neural networks, when applied to … like to thank the members of the Human Language Technology Group (Group …

Re-encoding in Neural Machine Translation
J Baptist – 2017 – esc.fnwi.uva.nl
… Better initialization strategies for deep neural networks such as the one proposed by [Glorot and Bengio, 2010] that make sure the weights of the neural network start in the right range depending on the type of activation functions and sizes of the network’s layers …

Retrieving Semantically Similar Questions in Community Question Answering
A Das – 2017 – pdfs.semanticscholar.org
… Answers” dataset reveals that DSTM and SCQA out- performs current state-of-the-art approaches based on translation models, topic models and existing deep neural network based models … 47 5.1.2 Exploring “parameter-sharing” in parallel legs of deep neural network …

Lemaza : An Arabic why-question answering system*
AM Azmi, NA Alshenaifi – Natural Language Engineering, 2017 – cambridge.org
… In a series of papers, the authors went on using different schemes to rank the candidate answers using supervised classifier (Support Vector Machine) (Severyn and Moschitti 2012), convolutional deep neural networks (Severyn and Moschitti 2015) and tree kernel (Tymoshenko …

Traducción automática basada en caracteres y redes neuronales
A LARRIBA FLOR – 2017 – riunet.upv.es
Page 1. Departamento de Sistemas Informáticos y Computación Universitat Politècnica de València Character-based Neural Machine Translation MASTER’S THESIS Master’s Degree in Artificial Intelligence, Pattern Recognition and Digital Imaging …

Role of Premises in Visual Question Answering
A Mahendru – 2017 – vtechworks.lib.vt.edu
Page 1. Role of Premises in Visual Question Answering Aroma Mahendru Thesis submitted to the Faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of Master of Science in Computer Engineering …

Self-training on refined clause patterns for relation extraction
DT Vo, E Bagheri – Information Processing & Management, 2017 – Elsevier
Skip to main content …

From Conditional Random Field (CRF) to Rhetorical Structure Theory (RST): Incorporating Context Information in Sentiment Analysis
A Vlachostergiou, G Marandianos, S Kollias – European Semantic Web …, 2017 – Springer
… to consider additional evaluation measures and finally to compare our proposed context incorporation method with deep neural network approaches … In: Proceedings of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp …

Manhattan distance
S Craw – Encyclopedia of Machine Learning and Data Mining, 2017 – Springer
… van den Driessche G, Schrittwieser J, Antonoglou I, Panneershelvam V, Lanctot M, Dieleman S, Grewe D, Nham J, Kalchbrenner N, Sutskever I, Lillicrap T, Leach M, Kavukcuoglu K, Graepel T, Hass- abis D (2016) Mastering the game of Go with deep neural networks and tree …

TalkBank and CLARIN
B MacWhinney – Selected papers from the CLARIN Annual Conference …, 2017 – ep.liu.se
Page 1. TalkBank and CLARIN Brian MacWhinney Department of Psychology Carnegie Mellon University, Pittsburgh USA macw@cmu.edu Abstract TalkBank promotes the use of corpora, web-based access, multimedia linkage …

Multimodal Crowdsourcing for Transcribing Handwritten Documents
E Granell, CD Martinez-Hinarejos – IEEE/ACM Transactions on …, 2017 – ieeexplore.ieee.org
Page 1. IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH, AND LANGUAGE PROCESSING, VOL. 25, NO. 2, FEBRUARY 2017 409 Multimodal Crowdsourcing for Transcribing Handwritten Documents Emilio Granell and Carlos-D. Mart?nez-Hinarejos …

EVALITA Goes Social: Tasks, Data
B Pierpaolo, N Malvina, S Rachele… – ITALIAN JOURNAL OF …, 2017 – iris.unito.it
Page 1. Volume 3, Number 1 june 2017 Emerging Topics at the Third Italian Conference on Computational Linguistics and EVALITA 2016 IJCoL Italian Journal Rivista Italiana of Computational Linguistics di Linguistica Computazionale ccademia university press aA Page …

TAMEEM V1. 0: speakers and text independent Arabic automatic continuous speech recognizer
MAM Abushariah – International Journal of Speech Technology, 2017 – Springer
… Accuracy rate. BC Broadcast conversations. BN Broadcast news. BR Broadcast reports. DBN Dynamic Bayesian networks. DNN Deep neural networks. HMM Hidden Markov model. HTK Hidden Markov model toolkit. LM Language model. LVQ Learning vector quantization. MFCC …

Sentiment-enhanced learning model for online language learning system
L Li – Electronic Commerce Research, 2017 – Springer
Page 1. Sentiment-enhanced learning model for online language learning system Li Li1 © Springer Science+Business Media, LLC, part of Springer Nature 2017 Abstract With the rapid development of the Internet, online learning …

Natural language processing for resource-poor languages
L Duong – 2017 – minerva-access.unimelb.edu.au
… We propose several algorithms motivated by transfer learning and incorpora- tion of additional resources in a low-resource setting. We also investigate an al- gorithm to model unwritten language based on deep neural networks. The more detailed contributions are as follows …

Constructing Sentences from Text Fragments: Aggregation in Text-to-text Generation
V Chenal – 2017 – digitool.library.mcgill.ca
Page 1. Constructing sentences from text fragments: Aggregation in text-to-text generation Victor Chenal Computer Science McGill University, Montreal August 7, 2017 A thesis submitted to McGill University in partial fulfilment of the requirements of the …

Modeling and Mining Domain Shared Knowledge for Sentiment Analysis
GY Zhou, JX Huang – ACM Transactions on Information Systems (TOIS), 2017 – dl.acm.org
… Recently, some efforts have been initiated on learning robust feature representations with deep neural networks (DNNs) in the context of sentiment classification. Socher et al … Zhou et al. (2016) proposed a bi-transferring deep neural networks framework for domain adaptation …

Turbo Decoders for Audio-visual Continuous Speech Recognition
AH Abdelaziz – Proc. Interspeech 2017, 2017 – pdfs.semanticscholar.org
… workshop on Human Language Technology. Association for Computational Linguistics, 1994, pp. 307–312. [27] K. Vesel`y, A. Ghoshal, L. Burget, and D. Povey, “Sequence- discriminative training of deep neural networks.” in Proc. Inter- speech, 2013, pp. 2345–2349 …

Named Entity Recognition with Word Embeddings and Wikipedia Categories for a Low-Resource Language
A Das, D Ganguly, U Garain – ACM Transactions on Asian and Low …, 2017 – dl.acm.org
Page 1. 18 Named Entity Recognition with Word Embeddings and Wikipedia Categories for a Low-Resource Language ARJUN DAS, University of Calcutta DEBASIS GANGULY, Dublin City University UTPAL GARAIN, Indian Statistical Institute …

Signal Processing Platforms and Algorithms for Real-life Communications and Listening to Digital Audio
A Petrovsky – downloads.hindawi.com
Page 1. Journal of Electrical and Computer Engineering Signal Processing Platforms and Algorithms for Real-life Communications and Listening to Digital Audio Lead Guest Editor: Alexander Petrovsky Guest Editors: Wanggen Wan, Manuel R. Zurera, and Alexey Karpov …

Advances on the Transcription of Historical Manuscripts based on Multimodality, Interactivity and Crowdsourcing
EG Romero – 2017 – riunet.upv.es
… CMN Cepstral Mean Normalisation. CN Confusion Network. CNC Confusion Network Combination. DFA Deterministic Finite Automaton. DNN Deep Neural Network. EFR E?ort Reduction. EM Expectation-Maximization. ER Error Rate. FST Finite State Transducer …

Acoustic Modeling of Under-Resourced Languages
R Sahraeian – 2017 – lirias.kuleuven.be
… CE Cross Entropy DD Data Driven DNN Deep Neural Network EM Expectation Maximization FBANK filter-bank … NCHLT National Center of Human Language Technology PDF Probability Density Function PER Phone Error Rate PLP Perceptual Linear Prediction …

Domain-specific sentiment classification via fusing sentiment knowledge from multiple sources
F Wu, Y Huang, Z Yuan – Information Fusion, 2017 – Elsevier
… They explored to use Stacked Denoising Autoencoders technique to learn generic concepts which are shared by different domains. These concepts are obtained from the hidden nodes of the deep neural network, and are used to construct a new feature space …

Inverted Alignments for End-to-End Automatic Speech Recognition
P Doetsch, M Hannemann, R Schlüter… – IEEE Journal of …, 2017 – ieeexplore.ieee.org
Page 1. IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING, VOL. 11, NO. 8, DECEMBER 2017 1265 Inverted Alignments for End-to-End Automatic Speech Recognition Patrick Doetsch , Mirko Hannemann, Ralf Schlüter, and Hermann Ney …

Relation Extraction: A Survey
S Pawar, GK Palshikar, P Bhattacharyya – arXiv preprint arXiv:1712.05191, 2017 – arxiv.org
Page 1. Relation Extraction : A Survey Sachin Pawara,b, Girish K. Palshikara, Pushpak Bhattacharyyab aTCS Research, Tata Consultancy Services Ltd. bDepartment of CSE, Indian Institute of Technology Bombay Abstract With …

Low latency acoustic modeling using temporal convolution and LSTMs
V Peddinti, Y Wang, D Povey… – IEEE Signal Processing …, 2017 – ieeexplore.ieee.org
… Daniel Povey and Sanjeev Khudanpur are with the Center for Language and Speech Processing(CLSP) and Human Language Technology Center of … O. Vinyals, A. Senior, and H. Sak, “Convolutional, long short-term memory, fully connected deep neural networks,” in Acoustics …

(Visited 99 times, 1 visits today)