Απεικόνιση αλγορίθµων Βαθιάς Μάθησης σε πλατφόρµες υλικού Graphcore IPU και NVIDIA GPU

Γερακάρης, Απόστολος; Gerakaris, Apostolos

dc.contributor.author	Γερακάρης, Απόστολος	el
dc.contributor.author	Gerakaris, Apostolos	en
dc.date.accessioned	2023-05-25T10:58:46Z
dc.date.available	2023-05-25T10:58:46Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/57774
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.25471
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/gr/	*
dc.subject	Όραση υπολογιστών	el
dc.subject	Ανίχνευση Προσώπων	el
dc.subject	Μηχανική Μάθηση	el
dc.subject	Εξαρτημένη Μάθηση	el
dc.subject	Face Detection	en
dc.subject	Ανίχνευση Οροσήμων	el
dc.subject	Machine Learning	en
dc.subject	Keras	en
dc.subject	Tensorflow	en
dc.subject	GPU	en
dc.subject	Eyeblink Conditioning	en
dc.subject	Landmark Detection	en
dc.title	Απεικόνιση αλγορίθµων Βαθιάς Μάθησης σε πλατφόρµες υλικού Graphcore IPU και NVIDIA GPU	el
heal.type	bachelorThesis
heal.generalDescription	Eυχαριστώ ιδιαίτερα τον ∆ρ. Χρήστο Στρύδη, επικεφαλή του Εργαστηρίου Νευροϋπολογιστών του Erasmus MC, για την καθοδήγησή του και την εξαιρετική συνεργασία που είχαµε.	el
heal.classification	Machine Learning	en
heal.classification	Μηχανική Μάθηση	el
heal.language	el
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2023-04-10
heal.abstract	This thesis investigates the performance of hardware accelerators, namely GPUs and IPUs, for Machine Learning and Deep Learning applications. Specifically, we focus on two tasks: automating and accelerating eyeblink-response detection from video and training an image-based CNN face detection model. For the former, we explore and compare different algorithms and optimization techniques to achieve real-time processing speed. In the latter, we optimize the training pipeline by leveraging both CPUs and device accelerators. The Eyeblink Conditioning experiment is a widely used experiment in the field of neuroscience to study learning and memory processes in the brain. In the past, researchers have used potentiometers or electromyography (EMG) to monitor the movement of the eyelid during an experiment. In recent years, the use of computer vision and image processing has greatly reduced the need for these methods, as they they need human intervention and do not allow real-time processing. In order to fully automate eyelid tracking, we chose a combination of face and landmark-detection algorithms and accelerated them to create a fast and accurate implementation. Various different algorithms from the fields of Deep Learning and Machine Learning are analyzed and compared for face detection and landmark detection (eyelid detection). Based on this study, two algorithms are identified as most suitable for our use case: the Ensemble of Regression Trees (ERT) approach for landmark detection and the BlazeFace CNN-based model for face detection. The BlazeFace model was accelerated on three different hardware accelerators: V100 Tesla GPU, MK1 IPU and MK2 IPU. The ERT algorithm was accelerated using multi-core CPUs. A combined implementation is successfully deployed for a real neuroscientific use-case: eyeblink response detection, achieving an overall runtime of 0.642 ms per frame with Tesla V100 GPU and 32 CPU processes, 0.7116 ms per frame with MK2-IPU and 64 CPU processes and 0.761 ms per frame with MK1-IPU and 32 CPU processes. Furthermore, an experimental open-source training implementation of the BlazeFace face detector was built from scratch to benchmark the performance of IPU and GPU hardware accelerators. Our results show that IPU-based systems have superior performance compared to the GPU-based systems in training the CNN-based face detector, especially for small batch sizes.	en
heal.advisorName	Πνευματικάτος, Διονύσιος	el
heal.committeeMemberName	Πνευματικάτος, Διονύσιος	el
heal.committeeMemberName	Κοζύρης, Νεκτάριος	el
heal.committeeMemberName	Γκούμας, Γεώργιος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών	el
heal.academicPublisherID	ntua
heal.fullTextAvailability	false