Search
Search results for: CONTINUAL LEARNING · MODEL MERGING
-
MagMax: Leveraging Model Merging for Seamless Continual Learning
PublicationThis paper introduces a continual learning approach named MagMax, which utilizes model merging to enable large pre-trained models to continuously learn from new data without forgetting previously acquired knowledge. Distinct from traditional continual learning methods that aim to reduce forgetting during task training, MagMax combines sequential fine-tuning with a maximum magnitude weight selection for effective knowledge integration...
-
Revisiting Supervision for Continual Representation Learning
Publication"In the field of continual learning, models are designed to learn tasks one after the other. While most research has centered on supervised continual learning, there is a growing interest in unsupervised continual learning, which makes use of the vast amounts of unlabeled data. Recent studies have highlighted the strengths of unsupervised methods, particularly self-supervised learning, in providing robust representations. The improved...
-
Divide and not forget: Ensemble of selectively trained experts in Continual Learning
PublicationClass-incremental learning is becoming more popular as it helps models widen their applicability while not forgetting what they already know. A trend in this area is to use a mixture-of-expert technique, where different models work together to solve the task. However, the experts are usually trained all at once using whole task data, which makes them all prone to forgetting and increasing computational burden. To address this limitation,...
-
Looking through the past: better knowledge retention for generative replay in continual learning
PublicationIn this work, we improve the generative replay in a continual learning setting to perform well on challenging scenarios. Because of the growing complexity of continual learning tasks, it is becoming more popular, to apply the generative replay technique in the feature space instead of image space. Nevertheless, such an approach does not come without limitations. In particular, we notice the degradation of the continually trained...
-
Adapt Your Teacher: Improving Knowledge Distillation for Exemplar-free Continual Learning
PublicationIn this work, we investigate exemplar-free class incremental learning (CIL) with knowledge distillation (KD) as a regularization strategy, aiming to prevent forgetting. KDbased methods are successfully used in CIL, but they often struggle to regularize the model without access to exemplars of the training data from previous tasks. Our analysis reveals that this issue originates from substantial representation shifts in the teacher...
-
Vident-synth: a synthetic intra-oral video dataset for optical flow estimation
Open Research DataWe introduce Vident-synth, a large dataset of synthetic dental videos with corresponding ground truth forward and backward optical flows and occlusion masks. It can be used for:
-
A continual model of a damaged medium used for analyzing fatigue life of polycrystalline structural alloys under thermal–mechanical loading
PublicationThe main physical laws of thermal–plastic deformation and fatigue damage accumulation processes in polycrystalline structural alloys under various regimes of cyclic thermal–mechanical loading are considered. Within the framework of mechanics of damaged media, a mathematical model is developed that describes thermal–plastic deformation and fatigue damage accumulation processes under low-cycle loading. The model consists of three...
-
Interpretable Deep Learning Model for the Detection and Reconstruction of Dysarthric Speech
PublicationWe present a novel deep learning model for the detection and reconstruction of dysarthric speech. We train the model with a multi-task learning technique to jointly solve dysarthria detection and speech reconstruction tasks. The model key feature is a low-dimensional latent space that is meant to encode the properties of dysarthric speech. It is commonly believed that neural networks are black boxes that solve problems but do not...
-
Measurement of the Development of a Learning IT Organization Supported by a Model of Knowledge Acquisition and Processing
PublicationThe paper presents a model of knowledge acquisition and processing for the development of learning organizations. The theory of a learning organization provides neither metrics nor tools to measure its development The authors' studies in this field are based on their experience gathered after projects realized in real IT organizations. The authors have described the construction of the model and the methods of its verification...
-
Model-free and Model-based Reinforcement Learning, the Intersection of Learning and Planning
PublicationMy doctoral dissertation is intended as the compound of four publications considering: structure and randomness in planning and reinforcement learning, continuous control with ensemble deep deterministic policy gradients, toddler-inspired active representation learning, and large-scale deep reinforcement learning costs.
-
Active Learning on Ensemble Machine-Learning Model to Retrofit Buildings Under Seismic Mainshock-Aftershock Sequence
PublicationThis research presents an efficient computational method for retrofitting of buildings by employing an active learning-based ensemble machine learning (AL-Ensemble ML) approach developed in OpenSees, Python and MATLAB. The results of the study shows that the AL-Ensemble ML model provides the most accurate estimations of interstory drift (ID) and residual interstory drift (RID) for steel structures using a dataset of 2-, to 9-story...
-
Bimodal deep learning model for subjectively enhanced emotion classification in films
PublicationThis research delves into the concept of color grading in film, focusing on how color influences the emotional response of the audience. The study commenced by recalling state-of-the-art works that process audio-video signals and associated emotions by machine learning. Then, assumptions of subjective tests for refining and validating an emotion model for assigning specific emotional labels to selected film excerpts were presented....
-
Model-Based Adaptive Machine Learning Approach in Concrete Mix Design
PublicationConcrete mix design is one of the most critical issues in concrete technology. This process aims to create a concrete mix which helps deliver concrete with desired features and quality. Contemporary requirements for concrete concern not only its structural properties, but also increasingly its production process and environmental friendliness, forcing concrete producers to use both chemically and technologically complex concrete...
-
A Proposed Machine Learning Model for Forecasting Impact of Traffic-Induced Vibrations on Buildings
PublicationTraffic-induced vibrations may cause various damages to buildings located near the road, including cracking of plaster, cracks in load-bearing elements or even collapse of the whole structure. Measurements of vibrations of real buildings are costly and laborious. Therefore the aim of the research is to propose the original numerical algorithm which allows us to predict, with high probability, the nega-tive dynamic impact of traffic-induced...
-
Fusion-based Representation Learning Model for Multimode User-generated Social Network Content
PublicationAs mobile networks and APPs are developed, user-generated content (UGC), which includes multi-source heterogeneous data like user reviews, tags, scores, images, and videos, has become an essential basis for improving the quality of personalized services. Due to the multi-source heterogeneous nature of the data, big data fusion offers both promise and drawbacks. With the rise of mobile networks and applications, UGC, which includes...
-
Optimized Deep Learning Model for Flood Detection Using Satellite Images
PublicationThe increasing amount of rain produces a number of issues in Kerala, particularly in urban regions where the drainage system is frequently unable to handle a significant amount of water in such a short duration. Meanwhile, standard flood detection results are inaccurate for complex phenomena and cannot handle enormous quantities of data. In order to overcome those drawbacks and enhance the outcomes of conventional flood detection...
-
Model szkolenia "Blended learning" z wykorzystaniem platformy Oracle I-learning.
PublicationW artykule zaproponowano modele organizacyjne szkoleń "blended learning", które pokazują możliwości współpracy firm prywatnych z instytucjami edukacyjnymi w dziedzinie e-learningu. W ramach wspólnego eksperymentu firm Oracle, Incenti S.A., WiedzaNet Sp. z o.o. oraz Centrum Edukacji Niestacjonarnej Politechniki Gdańskiej w semestrze letnim roku akademickiego 2003/2004 udostępniony będzie kurs dla studentów Wydziału Inzynierii Lądowej...
-
Projekt Leonardo da Vinci EMDEL (European Model for Distance Education and Learning) - otwarte szkolenia online.
PublicationW referacie zaprezentowano główne zadania oraz ofertę szkoleniową Centrum Edukacji Niestacjonarnej Politechniki Gdańskiej (CEN PG) w kontekście realizowanych projektów Unii Europejskiej. Przedstawiono projekt Leonardo da Vinci EMDEL - European Model for Distance Education and learning - realizowany przez CEN PG w latach 2001-2005 oraz opisano doświadczenia w zakresie adaptacji i lokalizacji opracowanych przez partnerów projektu...
-
Blended Learning Model for Computer Techniques for Students of Architecture
PublicationAbstract: The article summarizes two-year experience of implementing hybrid formula for teaching Computer Techniques at the Faculty of Architecture at the Gdansk University of Technology. Original educational e-materials, consisting of video clips, text and graphics instructions, as well as links to online resources are embedded in the university e-learning educational platform. The author discusses technical constraints associated...
-
Chained machine learning model for predicting load capacity and ductility of steel fiber–reinforced concrete beams
PublicationOne of the main issues associated with steel fiber–reinforced concrete (SFRC) beams is the ability to anticipate their flexural response. With a comprehensive grid search, several stacked models (i.e., chained, parallel) consisting of various machine learning (ML) algorithms and artificial neural networks (ANNs) were developed to predict the flexural response of SFRC beams. The flexural performance of SFRC beams under bending was...
-
Model of distributed learning objects repository for a heterogenic internet environment
PublicationW artykule wprowadzono pojęcie komponentu edukacyjnego jako rozszerzenie obiektu edukacyjnego o elementy zachowania (metody). Zaproponowane podejście jest zgodne z paradygmatem obiektowym. W oparciu o komponent edukacyjny zaprojektowano model budowy repozytorium materiałów edukacyjnych. Model ten jest oparty o usługi sieciowe i rejestry UDDI. Komponent edukacyjny oraz model repozytorium mogą znaleźć zastosowanie w konstrukcji zbiorów...
-
Application of a hybrid mechanistic/machine learning model for prediction of nitrous oxide (N2O) production in a nitrifying sequencing batch reactor
PublicationNitrous oxide (N2O) is a key parameter for evaluating the greenhouse gas emissions from wastewater treatment plants. In this study, a new method for predicting liquid N2O production during nitrification was developed based on a mechanistic model and machine learning (ML) algorithm. The mechanistic model was first used for simulation of two 15-day experimental trials in a nitrifying sequencing batch reactor. Then, model predictions...
-
Deep learning model for automated assessment of lexical stress of non-native english speakers
Publication -
Road traffic can be predicted by machine learning equally effectively as by complex microscopic model
PublicationSince high-quality real data acquired from selected road sections are not always available, a traffic control solution can use data from software traffic simulators working offline. The results show that in contrast to microscopic traffic simulation, the algorithms employing neural networks can work in real-time, so they can be used, among others, to determine the speed displayed on variable message road signs. This paper describes...
-
An Adaptive Network Model for a Double Bias Perspective on Learning from Mistakes within Organizations
PublicationAlthough making mistakes is a crucial part of learning, it is still often being avoided in companies as it is considered as a shameful incident. This goes hand in hand with a mindset of a boss who dominantly believes that mistakes usually have negative consequences and therefore avoids them by only accepting simple tasks. Thus, there is no mechanism to learn from mistakes. Employees working for and being influenced by such a boss...
-
Experimental investigations and prediction of WEDMed surface of nitinol SMA using SinGAN and DenseNet deep learning model
Publication -
A Computationally Efficient Model for Predicting Successful Memory Encoding Using Machine-Learning-based EEG Channel Selection
PublicationComputational cost is an important consideration for memory encoding prediction models that use data from dozens of implanted electrodes. We propose a method to reduce computational expense by selecting a subset of all the electrodes to build the prediction model. The electrodes were selected based on their likelihood of measuring brain activity useful for predicting memory encoding better than chance (in terms of AUC). A logistic...
-
An Adaptive Network Model Simulating the Effects of Different Culture Types and Leader Qualities on Mistake Handling and Organisational Learning
PublicationThis paper investigates computationally the following research hypotheses: (1) Higher flexibility and discretion in organizational culture results in better mistake management and thus better organizational learning, (2) Effective organizational learning requires a transformational leader to have both high social and formal status and consistency, and (3) Company culture and leader's behavior must align for the best learning effects....
-
An Adaptive Network Model Simulating the Effects of Different Culture Types and Leader Qualities on Mistake Handling and Organisational Learning
PublicationThis paper investigates computationally the following research hypotheses: (1) Higher flexibility and discretion in organizational culture results in better mistake management and thus better organizational learning, (2) Effective organizational learning requires a transformational leader to have both high social and formal status and consistency, and (3) Company culture and leader’s behavior must align for the best learning effects....
-
Personalized prediction of the secondary oocytes number after ovarian stimulation: A machine learning model based on clinical and genetic data
PublicationControlled ovarian stimulation is tailored to the patient based on clinical parameters but estimating the number of retrieved metaphase II (MII) oocytes is a challenge. Here, we have developed a model that takes advantage of the patient’s genetic and clinical characteristics simultaneously for predicting the stimulation outcome. Sequence variants in reproduction-related genes identified by next-generation sequencing were matched...
-
An automated learning model for twitter sentiment analysis using Ranger AdaBelief optimizer based Bidirectional Long Short Term Memory
PublicationSentiment analysis is an automated approach which is utilized in process of analysing textual data to describe public opinion. The sentiment analysis has major role in creating impact in the day-to-day life of individuals. However, a precise interpretation of text still relies as a major concern in classifying sentiment. So, this research introduced Bidirectional Long Short Term Memory with Ranger AdaBelief Optimizer (Bi-LSTM RAO)...
-
Rediscovering Automatic Detection of Stuttering and Its Subclasses through Machine Learning—The Impact of Changing Deep Model Architecture and Amount of Data in the Training Set
PublicationThis work deals with automatically detecting stuttering and its subclasses. An effective classification of stuttering along with its subclasses could find wide application in determining the severity of stuttering by speech therapists, preliminary patient diagnosis, and enabling communication with the previously mentioned voice assistants. The first part of this work provides an overview of examples of classical and deep learning...
-
Category Adaptation Meets Projected Distillation in Generalized Continual Category Discovery
Publication"Generalized Continual Category Discovery (GCCD) tackles learning from sequentially arriving, partially labeled datasets while uncovering new categories. Traditional methods depend on feature distillation to prevent forgetting the old knowledge. However, this strategy restricts the model’s ability to adapt and effectively distinguish new categories. To address this, we introduce a novel technique integrating a learnable projector...
-
Zastosowanie metody studium przypadku w kształceniu menedżerów
PublicationKształcenie z wykorzystaniem metod rozwiązywania problemów (problem-based learning) staje się coraz bardziej popularne na wszystkich poziomach kształcenia, również w edukacji biznesowej. Przykładem takiej metody jest studium przypadku (case study). Metoda studium przypadku pozwala na rozwijanie umiejętności i kompetencji wykorzystywanych przez menedżerów w ich pracy, np. umiejętności syntezy, identyfikacji problemów, czy podejmowania...
-
Music information retrieval—The impact of technology, crowdsourcing, big data, and the cloud in art.
PublicationThe exponential growth of computer processing power, cloud data storage, and crowdsourcing model of gathering data bring new possibilities to music information retrieval (mir) field. Mir is no longer music content retrieval only; the area also comprises the discovery of expressing feelings and emotions contained in music, incorporating other than hearing modalities for helping this issue, users’ profiling, merging music with social...
-
A Nonlinear Model of a Mesh Shell
PublicationFor a certain class of elastic lattice shells experiencing finite deformations, a continual model using the equations of the so-called six-parameter shell theory has been proposed. Within this model, the kinematics of the shell is described using six kinematically independent scalar degrees of freedom — the field of displacements and turns, as in the case of the Cosserat continuum, which gives reason to call the model under consideration...
-
Wiktoria Wojnicz dr hab. inż.
PeopleDSc in Mechanics (in the field of Biomechanics) - Lodz Univeristy of Technology, 2019 PhD in Mechanics (in the field of Biomechanics) - Lodz Univeristy of Technology, 2009 (with distinction) List of papers (2009 - ) Wojnicz W., Wittbrodt E., Analysis of muscles' behaviour. Part I. The computational model of muscle. Acta of Bioengineering and Biomechanics, Vol. 11, No.4, 2009, p. 15-21 Wojnicz W., Wittbrodt E., Analysis of...
-
Integration of thermographic data with the 3D object model
PublicationThe aim of the paper is to present new method for merging the 3D model data of the measured object with thermograms. Our technique is based on the combination of visual 3D imaging technique and thermal imaging technique, which maps the 2D thermograms on to 3D anatomical mesh model. The combination of these imaging modalities allows the generation of combined 3D and thermal data from which thermal signatures can be verified and...
-
Optymalizacja zasobów chmury obliczeniowej z wykorzystaniem inteligentnych agentów w zdalnym nauczaniu
PublicationRozprawa dotyczy optymalizacji zasobów chmury obliczeniowej, w której zastosowano inteligentne agenty w zdalnym nauczaniu. Zagadnienie jest istotne w edukacji, gdzie wykorzystuje się nowoczesne technologie, takie jak Internet Rzeczy, rozszerzoną i wirtualną rzeczywistość oraz deep learning w środowisku chmury obliczeniowej. Zagadnienie jest istotne również w sytuacji, gdy pandemia wymusza stosowanie zdalnego nauczania na dużą skalę...
-
On the Equations of the Surface Elasticity Model Based on the Theory of Polymeric Brushes
PublicationMotivating by theory of polymers, in particular, by the models of polymeric brushes we present here the homogenized (continual) two-dimensional (2D) model of surface elasticity. A polymeric brush consists of an system of almost aligned rigid polymeric chains. The interaction between chain links are described through Stockmayer potential, which take into account also dipole-dipole interactions. The presented 2D model can be treated...
-
Hossein Nejatbakhsh Esfahani Dr.
PeopleMy research interests lie primarily in the area of Learning-based Safety-Critical Control Systems, for which I leverage the following concepts and tools:-Robust/Optimal Control-Reinforcement Learning-Model Predictive Control-Data-Driven Control-Control Barrier Function-Risk-Averse Controland with applications to:-Aerial and Marine robotics (fixed-wing UAVs, autonomous ships and underwater vehicles)-Multi-Robot and Networked Control...
-
Federated Learning in Healthcare Industry: Mammography Case Study
PublicationThe paper focuses on the role of federated learning in a healthcare environment. The experimental setup involved different healthcare providers, each with their datasets. A comparison was made between training a deep learning model using traditional methods, where all the data is stored in one place, and using federated learning, where the data is distributed among the workers. The experiment aimed to identify possible challenges...
-
Computational Analysis of Transformational Organisational Change with Focus on Organisational Culture and Organisational Learning: An Adaptive Dynamical Systems Modeling Approach
PublicationTransformative Organisational Change becomes more and more significant both practically and academically, especially in the context of organisational culture and learning. However computational modeling and formalization of organisational change and learning processes are still largely unexplored. This chapter aims to provide an adaptive network model of transformative organisational change and translate a selection of organisational...
-
Adaptive Dynamical Systems Modelling of Transformational Organizational Change: with Focus on Organizational Culture and Organizational Learning
PublicationTransformative Organizational Change becomes more and more significant both practically and academically, especially in the context of organizational culture and learning. However computational modeling and a formalization of organizational change and learning processes are still largely unexplored. This paper aims to provide an adaptive network model of transformative organizational change and translate a selection of organizational...
-
Adaptive Dynamical Systems Modelling of Transformational Organizational Change with Focus on Organizational Culture and Organizational Learning
PublicationTransformative Organizational Change becomes more and more significant both practically and academically, especially in the context of organizational culture and learning. However computational modeling and a formalization of organizational change and learning processes are still largely unexplored. This paper aims to provide an adaptive network model of transformative organizational change and translate a selection of organizational...
-
Adjusted SpikeProp algorithm for recurrent spiking neural networks with LIF neurons
PublicationA problem related to the development of a supervised learning method for recurrent spiking neural networks is addressed in the paper. The widely used Leaky-Integrate-and-Fire model has been adopted as a spike neuron model. The proposed method is based on a known SpikeProp algorithm. In detail, the developed method enables gradient descent learning of recurrent or multi-layer feedforward spiking neural networks. The research included...
-
Active Control of Highly Autocorrelated Machinery Noise in Multivariate Nonminimum Phase Systems
PublicationIn this paper, a novel multivariate active noise control scheme, designed to attenuate disturbances with high autocorrelation characteristics and preserve background signals, is proposed. The algorithm belongs to the class of feedback controllers and, unlike the popular feedforward FX-LMS approach, does not require availability of a reference signal. The proposed approach draws its inspiration from the iterative learning control...
-
Adam Władziński
PeopleAdam Władziński, a PhD Candidate at Gdansk University of Technology, specializes in Biomedical Engineering with a focus on machine learning for image processing and blockchain technology. Holding a BEng and MSc in Electronics, Adam Władziński has developed a keen interest in applying advanced computational techniques to biological systems. During their master’s program, Adam Władziński explored laser spectroscopy, building a database...
-
A MODEL FOR FORECASTING PM10 LEVELS WITH THE USE OF ARTIFICIAL NEURAL NETWORKS
PublicationThis work presents a method of forecasting the level of PM10 with the use of artificial neural networks. Current level of particulate matter and meteorological data was taken into account in the construction of the model (checked the correlation of each variable and the future level of PM10), and unidirectional networks were used to implement it due to their ease of learning. Then, the configuration of the network (built on the...
-
Knowledge sharing and knowledge hiding in light of the mistakes acceptance component of learning culture- knowledge culture and human capital implications
PublicationPurpose: This study examines the micromechanisms of how knowledge culture fosters human capital development. Method: An empirical model was developed using the structural equation modeling method (SEM) based on a sample of 321 Polish knowledge workers employed in different industries. Findings: This study provides direct empirical evidence that tacit knowledge sharing supports human capital, whereas tacit knowledge hiding does...