Filters








1,927 Hits in 4.0 sec

Product Categorization with LSTMs and Balanced Pooling Views

Michael Skinner
2018 Annual International ACM SIGIR Conference on Research and Development in Information Retrieval  
We also demonstrate the positive impact of tightening the connections between recurrent and output layers through the use of pooling layers, and introduce the Balanced Pooling View architecture to take  ...  Recurrent Neural Networks (RNNs), and LSTMs in particular, have proven competitive in a wide variety of language modeling and classification tasks.  ...  Thank you as well to the fast.ai community for developing and supporting an amazing MOOC and deep learning library.  ... 
dblp:conf/sigir/Skinner18 fatcat:n5zw4dplafdwbidfsjbsh2y4e4

Fusion Techniques for Utterance-Level Emotion Recognition Combining Speech and Transcripts

Jilt Sebastian, Piero Pierucci
2019 Interspeech 2019  
We investigate the use of long short-term memory (LSTM) recurrent neural network (RNN) with pre-trained word embedding for text-based emotion recognition and convolutional neural network (CNN) with utterance-level  ...  Recent developments in multi-modal emotion recognition utilize deeplearning techniques to achieve remarkable performances, with models based on different features suitable for text, audio and vision.  ...  We also thank the members of Speech Technology Group at Telepathy Labs GmbH for their constructive feedback and reviews.  ... 
doi:10.21437/interspeech.2019-3201 dblp:conf/interspeech/SebastianP19 fatcat:ukldyq7r3nejjlchls27zqmwxy

Hierarchical Text Classification Using Dictionary Based Approach And Long-Short Term Memory

Karishma D. Shaikh, Amol C. Adamuthe
2018 Zenodo  
The text classification process has been well studied, but there are still many improvements in the classification and feature preparation, which can optimize the performance of classification for specific  ...  In this work, a novel unified model called LSTM is proposed for text categorization. II.  ...  In the paper, we implemented dictionary-based approach and LSTM for text classification. The dictionary-based approach is simple approach and in that, we used text preprocessing and dot product.  ... 
doi:10.5281/zenodo.1413306 fatcat:yi7fjmyufndqlhtcdtx4jawqca

Human Action Recognition Using Deep Learning Methods on Limited Sensory Data

Nilay Tufek, Murat Yalcin, Mucahit Altintas, Fatma Kalaoglu, Yi Li, Senem Kursun Bahadir
2019 IEEE Sensors Journal  
Data balancing and data augmentation methods were applied and accuracy rates were increased noticeably.  ...  Several deep learning methods like Convolutional Neural Network(CNN), Long-Short Term Memory (LSTM) with classical machine learning algorithms and their combinations were implemented and a performance  ...  Learning rate is determined 0.01 and optimization method is chosen Stochastic Gradient Decent with categorical cross-entropy loss function.  ... 
doi:10.1109/jsen.2019.2956901 fatcat:fqwwnu6mobhxhni7mt4ejpzt3y

A Stock Selection Method Based on Earning Yield Forecast Using Sequence Prediction Models [article]

Jessie Sun
2019 arXiv   pre-print
And then elaborates the workflow of stock pool creation, feature selection, data structuring, model setup and model evaluation.  ...  The LSTM and GRU models demonstrate superior performance of forecast accuracy over a traditional Feedforward Neural Network model. The GRU model slightly outperformed the LSTM model.  ...  In practice, it is suggested using a trial and error approach to choose between LSTM and GRU models for a specific problem with a specific dataset.  ... 
arXiv:1905.04842v1 fatcat:rjywpvluvvflhhotr72jerlvoi

Prediction is very hard, especially about conversion. Predicting user purchases from clickstream data in fashion e-commerce [article]

Luca Bigon, Giovanni Cassani, Ciro Greco, Lucas Lacasa, Mattia Pavoni, Andrea Polonioli, Jacopo Tagliabue
2019 arXiv   pre-print
and prepared a novel dataset of live shopping sessions from a major European e-commerce fashion website; second, we use the dataset to test in a controlled environment strong baselines and SOTA models  ...  However, due to the low frequency of conversion events and the noisiness of browsing data, classifying user sessions is very challenging.  ...  Jacopo Tagliabue wishes to thank Tooso's clients too, as they have been instrumental in the success of the company and have been always very receptive to the possibilities opened by A.I. in retail and  ... 
arXiv:1907.00400v1 fatcat:qjs2gqh6fff73pme2bqdlty7ty

Benchmark Performance of Machine And Deep Learning Based Methodologies for Urdu Text Document Classification [article]

Muhammad Nabeel Asim, Muhammad Usman Ghani, Muhammad Ali Ibrahim, Sheraz Ahmad, Waqar Mahmood, Andreas Dengel
2020 arXiv   pre-print
datasets CLE Urdu Digest 1000k, and CLE Urdu Digest 1Million with a significant margin of 32%, and 13% respectively.  ...  The source code and presented dataset are available at Github repository.  ...  The theory behind this model is the same as Chunting Zhou et al., [105] model except it used both LSTM and GRU along with max-pooling layers after CNN.  ... 
arXiv:2003.01345v1 fatcat:sxmaksohlvaodctpxeqjaenbvq

Enhance the Motion Cues for Face Anti-Spoofing using CNN-LSTM Architecture [article]

Xiaoguang Tu, Hengsheng Zhang, Mei Xie, Yao Luo, Yuefei Zhang, Zheng Ma
2019 arXiv   pre-print
Then we leverage Long Short-Term Memory (LSTM) with the extracted features as inputs to capture the temporal dynamics in videos.  ...  Experiments on Replay Attack and MSU-MFSD databases show that the proposed method yields state-of-the-art performance with better generalization ability compared with several other popular algorithms.  ...  A confusion loss layer based on LSTM loss and CNN loss is created to balance the learning level of CNN and LSTM. 3.1.  ... 
arXiv:1901.05635v1 fatcat:gxgxe4i7wvaarkw34xbpw2uisa

Transfer Learning for Sentiment Analysis Using BERT Based Supervised Fine-Tuning

Nusrat Jahan Prottasha, Abdullah As Sami, Md Kowsher, Saydul Akbar Murad, Anupam Kumar Bairagi, Mehedi Masud, Mohammed Baz
2022 Sensors  
Additionally, we explore various word embedding techniques, such as Word2Vec, GloVe, and fastText, and compare their performance to the BERT transfer learning strategy.  ...  The availability of these different worldviews and individuals' emotions empowers sentiment analysis.  ...  We worked with an unbalanced dataset. A well-balanced dataset improves efficiency significantly.  ... 
doi:10.3390/s22114157 pmid:35684778 pmcid:PMC9185586 fatcat:sydlxnbmxzeongbsuftbn2jhba

Network Traffic Anomaly Detection via Deep Learning

Konstantina Fotiadou, Terpsichori-Helen Velivassaki, Artemis Voulkidis, Dimitrios Skias, Sofia Tsekeridou, Theodore Zahariadis
2021 Information  
For this purpose, we exploit the Convolutional Neural Networks (CNNs), and the Long Short Term Memory Networks (LSTMs) in order to construct robust multi-class classifiers, able to assign each new network  ...  Complex threat patterns and malicious actors are able to cause severe damages to cyber-systems.  ...  In our scheme, we used 1D max-pooling with a pooling size of 2 and 1 stride size.  ... 
doi:10.3390/info12050215 fatcat:4wsbsakej5aajhhtx72zvr2m2a

Semantic Classification of Tabular Datasets via Character-Level Convolutional Neural Networks [article]

Paul Azunre, Craig Corcoran, Numa Dhamani, Jeffrey Gleason, Garrett Honke, David Sullivan, Rebecca Ruppel, Sandeep Verma, Jonathon Morgan
2019 arXiv   pre-print
In doing so, realistic data imperfections are learned and the set of classes handled can be expanded from the base set with reduced labeled data and computing power requirements.  ...  Results show the effectiveness and flexibility of this approach in three diverse domains: semantic classification of tabular data, age prediction from social media posts, and email spam classification.  ...  Views,  ... 
arXiv:1901.08456v1 fatcat:qel5qhfh45eopphyby3szwajmy

Short-Term Photovoltaic Power Forecasting Using a Convolutional Neural Network–Salp Swarm Algorithm

Happy Aprillia, Hong-Tzer Yang, Chao-Ming Huang
2020 Energies  
The results showed that the proposed CNN-SSA could accommodate the actual generation pattern better than the SVM-SSA and LSTM-SSA methods.  ...  To evaluate the performance of the proposed method, comparisons were made to the SSA based support vector machine (SVM-SSA) and long short-term memory neural network (LSTM-SSA) methods.  ...  production.  ... 
doi:10.3390/en13081879 fatcat:c5p73a4egncrzkl4qwqxdzwndu

Classification of Electrocardiography Hybrid Convolutional Neural Network-Long Short Term Memory with Fully Connected Layer

Dhanagopal Ramachandran, R. Suresh Kumar, Ahmed Alkhayyat, Rami Q. Malik, Prasanna Srinivasan, G. Guga Priya, Amsalu Gosu Adigo, Arpit Bhardwaj
2022 Computational Intelligence and Neuroscience  
The CNN, LSTM, and DNN algorithms are acceptable for viewing.  ...  The LSTM network and FCL additionally demonstrated that the unbalanced datasets associated with the ECG beat detection problem could be consistently resolved and that they were not susceptible to the accuracy  ...  We used LSTM with FCL to classify ECG arrhythmias that were out of balance. e categorization of imbalanced ECG data is frequently employed, according to studies [21, 43] .  ... 
doi:10.1155/2022/6348424 pmid:35860642 pmcid:PMC9293511 fatcat:du5vovlg2vhifjdsy7mhahyjyy

Medical Visual Question Answering: A Survey [article]

Zhihong Lin, Donghao Zhang, Qingyi Tac, Danli Shi, Gholamreza Haffari, Qi Wu, Mingguang He, Zongyuan Ge
2022 arXiv   pre-print
We summarize and discuss their techniques, innovation, and potential improvement. In the last part, we analyze some medical-specific challenges for the field and discuss future research directions.  ...  Given a medical image and a clinically relevant question in natural language, the medical VQA system is expected to predict a plausible and convincing answer.  ...  For example, in MIMIC-CXR [37] , a radiology report is often associated with two images of posteroanterior view and lateral view.  ... 
arXiv:2111.10056v2 fatcat:4dihtqmptbgj5lozrv3lfxqv7q

Multi-Scale Convolutional Recurrent Neural Network for Bearing Fault Detection in Noisy Manufacturing Environments

Seokju Oh, Seugmin Han, Jongpil Jeong
2021 Applied Sciences  
The failure of a facility to produce a product can have significant impacts on the quality of the product.  ...  vibration signals in the field, and the MS-CRNN inspects and classifies defects.  ...  Acknowledgments: This work was supported by the Smart Factory Technological R&D Program S2727115 funded by Ministry of SMEs and Startups (MSS, Korea).  ... 
doi:10.3390/app11093963 doaj:0b64fcf264484680b7e90b1e3256a12c fatcat:rscshqu4j5aixiz3bxazrpaycm
« Previous Showing results 1 — 15 out of 1,927 results