Papers by Μαρία Σαμαράκου
Deep learning is increasingly used in diverse application fields with results typically surpassin... more Deep learning is increasingly used in diverse application fields with results typically surpassing those of traditional machine learning techniques. The portfolio of available neural networks is wide, consisting of the full range in terms of complexity, from compact networks to large ones with multiple layers and parameters. This heterogeneity in the model topology is reflected, not necessarily linearly, on the required computational resources for training and inference. Similarly, the environments where the neural networks are trained and executed are transformed from fully-fledged centralized nodes to distributed architectures with constrained resources. In this view, computational resource requirements can be one of the criteria for resource usage management and neural network selection. In this work we measure the training times for a set of five convolutional neural networks of varying complexity and age (GoogleNet, ShuffleNet, VGGish, YAMNet) under different training configurations, considering the batch size, the number of epochs and the learning rate. These measurements are used to create a CPU-time-training dataset of more than 500 values. This dataset is used to train and evaluate models, based on neural networks, for estimating and predicting training times depending on the models employed and the training parameters. Five regression models have been trained and evaluated in terms of correlation coefficient and root mean square error. In addition, we measure the CPU times needed for inference for a subset of the trained models, which prove to be uncorrelated with the corresponding training times.
A rule based intelligent control for photovoltaic systems
A rule-based intelligent control system is proposed for a photovoltaic (PV) system which consists... more A rule-based intelligent control system is proposed for a photovoltaic (PV) system which consists of a PV generator, batteries, and variable load. The control system uses prediction to program the operation of the PV system. Specifically, the PV system is simulated on the basis of qualitative process theory. The intelligent system then allows for maximization of the life duration of the batteries by predicting and avoiding undesirable conditions as well as maximum possible coverage of the electrical demand based on the predicted available energy. An advantage of the control program is that it can enlarge its knowledge base (with the use of certain additional rules) by incorporating real data. Changing the contents of the knowledge base makes it possible to modify the performance of the controller so that it can be used in different systems with different needs
CNN-based, contextualized, real-time fire detection in computational resource-constrained environments
Energy Reports
Expert Systems With Applications, May 1, 2007
In this paper, the implementation of a neural network-based fuzzy modeling approach to assess asp... more In this paper, the implementation of a neural network-based fuzzy modeling approach to assess aspects of students' learning style in the discovery learning environment ''Vectors in Physics and Mathematics'' is presented. Fuzzy logic is used to provide a linguistic description of students' behavior and learning characteristics, as they have been elicited from teachers, and to handle the inherent uncertainty associated with teachers' subjective assessments. Neural networks are used to add learning and generalization abilities to the fuzzy model by encoding teachers' experience through supervised neural-network learning. The neural network-based fuzzy diagnostic model is a general diagnostic model which is implemented in an Intelligent Learning Environment by eliciting teachers' expertise regarding students' characteristics based on real students' observation and on data being collected from students' interaction. The model has been successfully implemented, trained and tested in the learning environment ''Vectors in Physics and Mathematics'' by using the recommendations of a group of five experienced teachers. The performance of our model in real classroom conditions has been evaluated during an experiment with an experienced Physics teacher and 49 students of secondary school attending Physics lessons.
Proceedings of the 26th Pan-Hellenic Conference on Informatics
Decarbonization of energy generation is required to address the urgent need to combat climate cha... more Decarbonization of energy generation is required to address the urgent need to combat climate change and meet the Paris Agreement targets. Photovoltaic Systems (PVS) are a major contributor to this solution, and they are expected to grow significantly faster than any other renewable-energy technology in the coming years due to their appealing features. This paper discusses the development of a two-way communication protocol between two transceivers and a custom-designed communication board installed on each PV array. With this configuration, it is possible to transmit the measurements of each PV cell in the array to a data recording and monitoring service. Additionally, the application can customize the operation of the strings that make up the array, by broadcasting special commands to them. A prototype experimental setup was constructed in order to test the communication between the PV array communication board and the uplink/downlink transceivers. The experimental results verify the feasibility and effectiveness of the proposed communication protocol between the Electronics Interface System (EIS) and the Smart PV array since the communication protocol meets the specification for transmitting the information from 1000 smart PV cells within one minute, as well as the strict timing requirements by the protocol for transmitting the parameters to the PV strings. CCS CONCEPTS • Hardware → Printed circuit boards; PCB design and layout; Communication hardware, interfaces and storage; Power and energy; Integrated circuits.
An e-Learning System for Extracting Text Comprehension and Learning Style Characteristics
Educational Technology & Society, 2018
Technology mediated learning is very actively and widely researched, with numerous e-learning env... more Technology mediated learning is very actively and widely researched, with numerous e-learning environments designed for different educational purposes developed during the past few decades. Still, their organization and texts are not structured according to any theory of educational comprehension. Modern education is even more flexible and, thus, demanding, requiring the combination of multiple educational theories for effective results. In this paper we present the combination of two educational theories for text comprehension and learning styles, that are in use by the newly developed Student Diagnosis, Assistance, Evaluation System based on Artificial Intelligence (StuDiAsE) an open learning system for unattended student diagnosis, assistance and evaluation based on artificial intelligence. Α trial test about the role of learning styles in student profiling for text comprehension in the educational environment StuDiAsE is described. Research was run with participation of students using the environment for prior knowledge test and text activities. The process revealed remarkable results about the role of learning styles in students’ profiles for text comprehension. Three dimensions for learning styles were identified: conceptualization, visualization and progression dimension and were used for profiling. Refinement of profiles incorporates learning styles by decoding student behavior, which reflects student learning styles for text comprehension.
ΣΥΜΒΟΛΗ ΣΤΗ ΜΕΛΕΤΗ ΟΡΘΟΛΟΓΙΣΜΟΥ ΣΥΣΤΗΜΑΤΟΣ ΠΑΡΑΓΩΓΗΣ ΗΛΕΚΤΡΙΚΗΣ ΕΝΕΡΓΕΙΑΣ ΑΠΟ ΣΥΝΔΥΑΣΜΕΝΗ ΧΡΗΣΗ ΗΠΙΩΝ ΚΑΙ ΚΛΑΣΣΙΚΩΝ ΠΗΓΩΝ
Βελτιστοποίηση ενεργειακών συστημάτων
Information Sciences, Feb 25, 2005
In this paper, a neural network implementation for a fuzzy logic-based model of the diagnostic pr... more In this paper, a neural network implementation for a fuzzy logic-based model of the diagnostic process is proposed as a means to achieve accurate student diagnosis and updates of the student model in Intelligent Learning Environments. The neuro-fuzzy synergy allows the diagnostic model to some extent "imitate" teachers in diagnosing students' characteristics, and equips the intelligent learning environment with reasoning capabilities that can be further user to drive pedagogical decisions depending on the student learning style. The neuro-fuzzy implementation helps to encode both structured and non-structured teachers' knowledge: when teachers' reasoning is available and well defined, it can be encoded in the form of fuzzy rules; when teachers' reasoning is not well defined but is available through practical examples illustrating their experience, then the networks can be trained to represent this experience. The proposed approach has been tested in diagnosing aspects of student's learning style in a discovery-learning environment that aims to help students to construct the concepts of vectors in physics and mathematics. The diagnosis outcomes of the model have been compared against the recommendations of a group of five experienced teachers, and the results produced by two alternative soft computing methods. The results of our pilot study show that the neuro-fuzzy model successfully manages the inherent uncertainty of the diagnostic process; especially for marginal cases, i.e. where it is very difficult, even for human tutors, to diagnose and accurately evaluate students by directly synthesizing subjective and, some times, conflicting judgments.
ΣΥΜΒΟΛΗ ΣΤΗ ΜΕΛΕΤΗ ΟΡΘΟΛΟΓΙΣΜΟΥ ΣΥΣΤΗΜΑΤΟΣ ΠΑΡΑΓΩΓΗΣ ΗΛΕΚΤΡΙΚΗΣ ΕΝΕΡΓΕΙΑΣ ΑΠΟ ΣΥΝΔΥΑΣΜΕΝΗ ΧΡΗΣΗ ΗΠΙΩΝ ΚΑΙ ΚΛΑΣΣΙΚΩΝ ΠΗΓΩΝ
An analog computer system for the evaluation of hip joint moments during normal walking
A semantically-enriched, open environment for multimedia content delivery
Lecture Notes in Computer Science, 2009
In this paper we discuss recent directions concerning the structural analysis of science document... more In this paper we discuss recent directions concerning the structural analysis of science documents and cognitive aspects of document elements aiming at document comprehension. Structural analysis of documents, according to text comprehension theory, promotes document understanding and enhances universal accessibility of documents. We outline the process of structuring science documents and activities for comprehension using the authoring tool Re-TuDiSAuth. This process improves universal accessibility of document supporting authors to structure text and text activities for students with different abilities, requirements and preferences.
Efficient use of solar energy for commercial DC compressors for solarapplications
International Journal of Emerging Technologies in Learning (iJET), 2014
The personalized diagnosis, assistance and evaluation of students in open learning environments c... more The personalized diagnosis, assistance and evaluation of students in open learning environments can be a challenging task, especially in cases that the processes need to be taking place in real-time, classroom conditions. This paper describes the design of an open learning environment under development, designed to monitor the comprehension of students, assess their prior knowledge, build individual learner profiles, provide personalized assistance and, finally, evaluate their performance by using artificial intelligence. A trial test has been performed, with the participation of 20 students, which displayed promising results.
Μοντέλο αναπαράστασης προγράμματος ύλης μέσα σε ένα προσαρμοστικό εκπαιδευτικό λογισμικό
In this paper, we propose a framework for the design of learning activities, which follows the ke... more In this paper, we propose a framework for the design of learning activities, which follows the key idea of the "Learning-for-Use" model and adopts characteristics from the exploratory and collaborative learning. We elaborated on the framework in the design of learning activities for Introductory Programming courses and we developed and evaluated in a real classroom environment, learning activities for the teaching of the infinite "while" loop. The experimental results are encouraging for the effectiveness of the activities in enhancing learning and indicate that the framework enables the students to gain deeper understanding of the programming constructs under consideration and to apply them more effectively.
Effect of reactor temperature on energy efficiency of biogas plants
Έξυπνο σύστημα εκτίμησης μαθησιακών ικανοτήτων μαθητών δημοτικού με χαμηλή απόδοση στην επίλυση απλών λεκτικών προβλημάτων
Modern scientific advances have enabled remarkable efficacy for photovoltaic systems with regard ... more Modern scientific advances have enabled remarkable efficacy for photovoltaic systems with regard to the exploitation of solar energy, boosting them into having a rapidly growing position among the systems developed for the production of renewable energy. However, in many cases the design, analysis, and control of photovoltaic systems are tasks which are quite complex and thus difficult to be carried out. In order to cope with this kind of problems, appropriate software tools have been developed either as standalone products or parts of general purpose software platforms used to model and simulate the generation, transmission, and distribution of solar energy. The utilization of this kind of software tools may be extremely helpful to the successful performance evaluation of energy systems with maximum accuracy and minimum cost in time and effort. The work presented in this paper aims on a first level at the performance analysis of various configurations of photovoltaic systems through computer-aided modelling. On a second level, it provides a comparative evaluation of the credibility of two of the most advanced graphical programming environments, namely, Simulink and LabVIEW, with regard to their application in photovoltaic systems.
Uploads
Papers by Μαρία Σαμαράκου