dc.contributor.author | Agathocleous, Michalis | en |
dc.contributor.author | Christodoulou, Chris C. | en |
dc.contributor.author | Promponas, Vasilis J. | en |
dc.contributor.author | Kountouris, P. | en |
dc.contributor.author | Vassiliades, Vassilis | en |
dc.contributor.editor | Villa A.E.P. | en |
dc.contributor.editor | Masulli P. | en |
dc.contributor.editor | Rivero A.J.P. | en |
dc.creator | Agathocleous, Michalis | en |
dc.creator | Christodoulou, Chris C. | en |
dc.creator | Promponas, Vasilis J. | en |
dc.creator | Kountouris, P. | en |
dc.creator | Vassiliades, Vassilis | en |
dc.date.accessioned | 2019-11-13T10:38:11Z | |
dc.date.available | 2019-11-13T10:38:11Z | |
dc.date.issued | 2016 | |
dc.identifier.issn | 0302-9743 | |
dc.identifier.uri | http://gnosis.library.ucy.ac.cy/handle/7/53500 | |
dc.description.abstract | Predictions on sequential data, when both the upstream and downstream information is important, is a difficult and challenging task. The Bidirectional Recurrent Neural Network (BRNN) architecture has been designed to deal with this class of problems. In this paper, we present the development and implementation of the Scaled Conjugate Gradient (SCG) learning algorithm for BRNN architectures. The model has been tested on the Protein Secondary Structure Prediction (PSSP) and Transmembrane Protein Topology Prediction problems (TMPTP). Our method currently achieves preliminary results close to 73% correct predictions for the PSSP problem and close to 79% for the TMPTP problem, which are expected to increase with larger datasets, external rules, ensemble methods and filtering techniques. Importantly, the SCG algorithm is training the BRNN architecture approximately 3 times faster than the Backpropagation Through Time (BPTT) algorithm. © Springer International Publishing Switzerland 2016. | en |
dc.source | 25th International Conference on Artificial Neural Networks, ICANN 2016 | en |
dc.source.uri | https://www.scopus.com/inward/record.uri?eid=2-s2.0-84987935043&doi=10.1007%2f978-3-319-44778-0_15&partnerID=40&md5=f8ec536084fbdf36442a125c4390cc04 | |
dc.subject | Learning algorithms | en |
dc.subject | Artificial intelligence | en |
dc.subject | Forecasting | en |
dc.subject | Neural networks | en |
dc.subject | Topology | en |
dc.subject | Network architecture | en |
dc.subject | Recurrent neural networks | en |
dc.subject | Backpropagation algorithms | en |
dc.subject | Bioinformatics | en |
dc.subject | Proteins | en |
dc.subject | Conjugate gradient method | en |
dc.subject | Learning systems | en |
dc.subject | Backpropagation through time algorithms | en |
dc.subject | Bidirectional recurrent neural networks | en |
dc.subject | Computational intelligence | en |
dc.subject | Ensemble methods | en |
dc.subject | Filtering technique | en |
dc.subject | Protein secondary structure prediction | en |
dc.subject | Protein secondary-structure prediction | en |
dc.subject | Scaled conjugate gradient | en |
dc.subject | Scaled conjugate gradient algorithm | en |
dc.subject | Scaled conjugate gradients | en |
dc.subject | Trans-membrane proteins | en |
dc.subject | Transmembrane protein topology prediction | en |
dc.title | Training bidirectional recurrent neural network architectures with the scaled conjugate gradient algorithm | en |
dc.type | info:eu-repo/semantics/article | |
dc.identifier.doi | 10.1007/978-3-319-44778-0_15 | |
dc.description.volume | 9886 LNCS | en |
dc.description.startingpage | 123 | |
dc.description.endingpage | 131 | |
dc.author.faculty | 002 Σχολή Θετικών και Εφαρμοσμένων Επιστημών / Faculty of Pure and Applied Sciences | |
dc.author.department | Τμήμα Πληροφορικής / Department of Computer Science | |
dc.type.uhtype | Article | en |
dc.description.notes | <p>Sponsors: | en |
dc.description.notes | Conference code: 180929</p> | en |
dc.source.abbreviation | Lect. Notes Comput. Sci. | en |
dc.contributor.orcid | Promponas, Vasilis J. [0000-0003-3352-4831] | |
dc.contributor.orcid | Christodoulou, Chris C. [0000-0001-9398-5256] | |
dc.gnosis.orcid | 0000-0003-3352-4831 | |
dc.gnosis.orcid | 0000-0001-9398-5256 | |