Academic literature on the topic 'Hybrid deep learning (DL) model'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the lists of relevant articles, books, theses, conference reports, and other scholarly sources on the topic 'Hybrid deep learning (DL) model.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Journal articles on the topic "Hybrid deep learning (DL) model"
1
Tolba, Ahmed, Nihal N. Mostafa, and Karam M. Sallam. "Hybrid Deep Learning-Based Model for Intrusion Detection." Artificial Intelligence in Cybersecurity 1 (January 11, 2024): 1–11. http://dx.doi.org/10.61356/j.aics.2024.1198.
Full textAbstract:
There is an intensive need for intrusion detection systems (IDSs) due to incremental and frequent cyber-attacks. The first line of defense against online threats is an IDS. Researchers are using deep learning (DL) approaches to detect attackers and preserve user information. In this study, we introduce a hybrid DL-based model. The proposed model integrates LSTM and ResNet to eliminate the vanishing gradient problem and increase the accuracy of the classification model. The proposed model aims to classify between normal or an attack, with each attack either being a DoS, U2R, R2L, or a probe over the NSL-KDD dataset. The proposed model achieves 99.5% according to accuracy. The model was compared with other ML and DL models.
2
Flores, Nahum, José La Rosa, Sebastian Tuesta, Luis Izquierdo, María Henriquez, and David Mauricio. "Hybrid Deep Learning Model for Improved Glaucoma Diagnostic Accuracy." Information 16, no. 7 (2025): 593. https://doi.org/10.3390/info16070593.
Full textAbstract:
Glaucoma is an irreversible neurodegenerative disease that affects the optic nerve, leading to partial or complete vision loss. Early and accurate detection is crucial to prevent vision impairment, which necessitates the development of highly precise diagnostic tools. Deep learning (DL) has emerged as a promising approach for glaucoma diagnosis, where the model is trained on datasets of fundus images. To improve the detection accuracy, we propose a hybrid model for glaucoma detection that combines multiple DL models with two fine-tuning strategies and uses a majority voting scheme to determine the final prediction. In experiments, the hybrid model achieved a detection accuracy of 96.55%, a sensitivity of 98.84%, and a specificity of 94.32%. Integrating datasets was found to improve the performance compared to using them separately even with transfer learning. When compared to individual DL models, the hybrid model achieved a 20.69% improvement in accuracy compared to the best model when applied to a single dataset, a 13.22% improvement when applied with transfer learning across all datasets, and a 1.72% improvement when applied to all datasets. These results demonstrate the potential of hybrid DL models to detect glaucoma more accurately than individual models.
3
T, S. Chandrakantha, N. Jagadale Basavaraj, and Abdullah Murshed Farhan Alnaggar Omar. "Lung Tumor Classification using Hybrid Deep Learning and Segmentation by Fuzzy C Means." Indian Journal of Science and Technology 17, no. 1 (2024): 70–79. https://doi.org/10.17485/IJST/v17i1.2124.
Full textAbstract:
Abstract <strong>Objectives:</strong> This study aims to employ a hybrid Deep Learning (DL) technique for automating tumor detection and classification in lung scans.<strong> Methods:</strong> The methodology involves three key stages: data preparation, segmentation using Fuzzy C Means (FCM), and classification using a hybrid DL model. The image dataset is sourced from the benchmark Lung Tumor (LT) data, and for segmentation, the FCM approach is applied. The hybrid DL model is created by combining a Pulse Coupled Neural Network (PCNN) and a Convolutional Neural Network (CNN). The study utilizes a dataset of 300 individuals from the NSCLC-Radiomics database. The validation process employs DICE and sensitivity for segmentation, while the hybrid model's confusion matrix elements contribute to performance validation. FCM and the hybrid model are employed for processing, segmenting, and classifying the images. Evaluation metrics such as Dice similarity and Sensitivity gauge the success of the segmentation method by measuring the intersection between ground truths and predictions. After segmentation evaluation, the classification process is executed, employing accuracy and loss in the training phase and metrics like accuracy and F1-score in the testing phase for model validation. <strong>Findings:</strong> The proposed approach achieves an accuracy of 97.43% and an F1-score of 98.28%. These results demonstrate the effectiveness of the suggested approach in accurately classifying and segmenting lung tumors. <strong>Novelty:</strong> The primary contribution of the research is a hybrid DL model based on PCCN+CCN. This ultimately raises the quality of the model, and these are carried out using real-time public medical images, demonstrating the model's originality. <strong>Keywords:</strong> Lung, Tumor, Segmentation, Classification, Hybrid model
4
Choudhary, Laxmi, and Jitendra Singh Choudhary. "Deep Learning Meets Machine Learning: A Synergistic Approach towards Artificial Intelligence." Journal of Scientific Research and Reports 30, no. 11 (2024): 865–75. http://dx.doi.org/10.9734/jsrr/2024/v30i112614.
Full textAbstract:
The evolution of artificial intelligence (AI) has progressed from rule-based systems to learning-based models, integrating machine learning (ML) and deep learning (DL) to tackle complex data-driven tasks. This review examines the synergy between ML, which utilizes algorithms like decision trees and support vector machines for structured data, and DL, which employs neural networks for processing unstructured data such as images and natural language. The combination of these paradigms through hybrid ML-DL models has enhanced prediction accuracy, scalability, and automation across domains like healthcare, finance, natural language processing, and robotics. However, challenges such as computational demands, data dependency, and model interpretability remain. This paper discusses the benefits, limitations, and future potential of ML and DL and also provides a review study of a hybrid model makes use of both techniques (machine learning & deep learning) advantages to solve complicated problems more successfully than one could on its own. To boost performance, increase efficiency, or address scenarios where either ML or DL alone would not be able to manage, this approach combines deep learning structures with conventional machine learning techniques.
5
Madhuri Ghuge, Et al. "Deep Learning Driven QoS Anomaly Detection for Network Performance Optimization." Journal of Electrical Systems 19, no. 2 (2024): 97–104. http://dx.doi.org/10.52783/jes.695.
Full textAbstract:
In modern, ever-changing network environments, QoS must be high to provide reliable and efficient services. This study tests Deep Learning (DL), specifically CNN, LSTM, and a hybrid CNN-LSTM model, to identify abnormalities using QoS measurements like Availability, Bandwidth, Latency, Jitter, and Packet Loss. The study evaluates DL-based QoS management using UNSW-NB15 data. The hybrid CNN-LSTM model excels at QoS management, identifying anomalies in key metrics with few false detections. This method captures intricate network data patterns and interrelationships using deep learning, improving anomaly detection accuracy and efficiency. A hybrid model is used to quantify QoS parameters like Availability, Bandwidth, Latency, Jitter, and Packet Loss. The results show high values for Packet Delivery Ratio (PDR), Throughput, Round-Trip Time (RTT), Variation in RTT, and Packet Loss Rate (PLR), proving the proposed approach's effectiveness in maintaining QoS. The CNN, LSTM, and suggested hybrid model evaluation metrics include Accuracy, Precision, Recall, and F1-Score. The hybrid model outperforms the individual models with 98.67% accuracy, precision, recall, and F1-Score. This proves its anomaly detection resilience. False Positive Rate and True Positive Rate show that the hybrid model performs best, with a 0.01 false positive rate and 0.98 true positive rate. Graphical representations help visualize DL model parameter comparisons and False Positive/True Positive rates. DL-based methods, particularly the hybrid CNN-LSTM model, are crucial for QoS anomaly detection in this study. Measurable results show the model improves network dependability, resource allocation, and user satisfaction. The study also suggests researching advanced deep learning methods for real-time network processing and scalable solutions.
6
Quan, Changqin, Zhiwei Luo, and Song Wang. "A Hybrid Deep Learning Model for Protein–Protein Interactions Extraction from Biomedical Literature." Applied Sciences 10, no. 8 (2020): 2690. http://dx.doi.org/10.3390/app10082690.
Full textAbstract:
The exponentially increasing size of biomedical literature and the limited ability of manual curators to discover protein–protein interactions (PPIs) in text has led to delays in keeping PPI databases updated with the current findings. The state-of-the-art text mining methods for PPI extraction are primarily based on deep learning (DL) models, and the performance of a DL-based method is mainly affected by the architecture of DL models and the feature embedding methods. In this study, we compared different architectures of DL models, including convolutional neural networks (CNN), long short-term memory (LSTM), and hybrid models, and proposed a hybrid architecture of a bidirectional LSTM+CNN model for PPI extraction. Pretrained word embedding and shortest dependency path (SDP) embedding are fed into a two-embedding channel model, such that the model is able to model long-distance contextual information and can capture the local features and structure information effectively. The experimental results showed that the proposed model is superior to the non-hybrid DL models, and the hybrid CNN+Bidirectional LSTM model works well for PPI extraction. The visualization and comparison of the hidden features learned by different DL models further confirmed the effectiveness of the proposed model.
7
Wang, Xiaomei, Ijaz Ahmad, Danish Javeed, et al. "Intelligent Hybrid Deep Learning Model for Breast Cancer Detection." Electronics 11, no. 17 (2022): 2767. http://dx.doi.org/10.3390/electronics11172767.
Full textAbstract:
Breast cancer (BC) is a type of tumor that develops in the breast cells and is one of the most common cancers in women. Women are also at risk from BC, the second most life-threatening disease after lung cancer. The early diagnosis and classification of BC are very important. Furthermore, manual detection is time-consuming, laborious work, and, possibility of pathologist errors, and incorrect classification. To address the above highlighted issues, this paper presents a hybrid deep learning (CNN-GRU) model for the automatic detection of BC-IDC (+,−) using whole slide images (WSIs) of the well-known PCam Kaggle dataset. In this research, the proposed model used different layers of architectures of CNNs and GRU to detect breast IDC (+,−) cancer. The validation tests for quantitative results were carried out using each performance measure (accuracy (Acc), precision (Prec), sensitivity (Sens), specificity (Spec), AUC and F1-Score. The proposed model shows the best performance measures (accuracy 86.21%, precision 85.50%, sensitivity 85.60%, specificity 84.71%, F1-score 88%, while AUC 0.89 which overcomes the pathologist’s error and miss classification problem. Additionally, the efficiency of the proposed hybrid model was tested and compared with CNN-BiLSTM, CNN-LSTM, and current machine learning and deep learning (ML/DL) models, which indicated that the proposed hybrid model is more robust than recent ML/DL approaches.
8
Li, Lu, Yongjiu Dai, Zhongwang Wei, et al. "Enforcing Water Balance in Multitask Deep Learning Models for Hydrological Forecasting." Journal of Hydrometeorology 25, no. 1 (2024): 89–103. http://dx.doi.org/10.1175/jhm-d-23-0073.1.
Full textAbstract:
Abstract Accurate prediction of hydrological variables (HVs) is critical for understanding hydrological processes. Deep learning (DL) models have shown excellent forecasting abilities for different HVs. However, most DL models typically predicted HVs independently, without satisfying the principle of water balance. This missed the interactions between different HVs in the hydrological system and the underlying physical rules. In this study, we developed a DL model based on multitask learning and hybrid physically constrained schemes to simultaneously forecast soil moisture, evapotranspiration, and runoff. The models were trained using ERA5-Land data, which have water budget closure. We thoroughly assessed the advantages of the multitask framework and the proposed constrained schemes. Results showed that multitask models with different loss-weighted strategies produced comparable or better performance compared to the single-task model. The multitask model with a scaling factor of 5 achieved the best among all multitask models and performed better than the single-task model over 70.5% of grids. In addition, the hybrid constrained scheme took advantage of both soft and hard constrained models, providing physically consistent predictions with better model performance. The hybrid constrained models performed the best among different constrained models in terms of both general and extreme performance. Moreover, the hybrid model was affected the least as the training data were artificially reduced, and provided better spatiotemporal extrapolation ability under different artificial prediction challenges. These findings suggest that the hybrid model provides better performance compared to previously reported constrained models when facing limited training data and extrapolation challenges.
9
Saleh, Hager, Sherif Mostafa, Lubna Abdelkareim Gabralla, Ahmad O. Aseeri, and Shaker El-Sappagh. "Enhanced Arabic Sentiment Analysis Using a Novel Stacking Ensemble of Hybrid and Deep Learning Models." Applied Sciences 12, no. 18 (2022): 8967. http://dx.doi.org/10.3390/app12188967.
Full textAbstract:
Sentiment analysis (SA) is a machine learning application that drives people’s opinions from text using natural language processing (NLP) techniques. Implementing Arabic SA is challenging for many reasons, including equivocation, numerous dialects, lack of resources, morphological diversity, lack of contextual information, and hiding of sentiment terms in the implicit text. Deep learning models such as convolutional neural networks (CNN) and long short-term memory (LSTM) have significantly improved in the Arabic SA domain. Hybrid models based on CNN combined with long short-term memory (LSTM) or gated recurrent unit (GRU) have further improved the performance of single DL models. In addition, the ensemble of deep learning models, especially stacking ensembles, is expected to increase the robustness and accuracy of the previous DL models. In this paper, we proposed a stacking ensemble model that combined the prediction power of CNN and hybrid deep learning models to predict Arabic sentiment accurately. The stacking ensemble algorithm has two main phases. Three DL models were optimized in the first phase, including deep CNN, hybrid CNN-LSTM, and hybrid CNN-GRU. In the second phase, these three separate pre-trained models’ outputs were integrated with a support vector machine (SVM) meta-learner. To extract features for DL models, the continuous bag of words (CBOW) and the skip-gram models with 300 dimensions of the word embedding were used. Arabic health services datasets (Main-AHS and Sub-AHS) and the Arabic sentiment tweets dataset were used to train and test the models (ASTD). A number of well-known deep learning models, including DeepCNN, hybrid CNN-LSTM, hybrid CNN-GRU, and conventional ML algorithms, have been used to compare the performance of the proposed ensemble model. We discovered that the proposed deep stacking model achieved the best performance compared to the previous models. Based on the CBOW word embedding, the proposed model achieved the highest accuracy of 92.12%, 95.81%, and 81.4% for Main-AHS, Sub-AHS, and ASTD datasets, respectively.
10
Sagu, Amit, Nasib Singh Gill, Preeti Gulia, Pradeep Kumar Singh, and Wei-Chiang Hong. "Design of Metaheuristic Optimization Algorithms for Deep Learning Model for Secure IoT Environment." Sustainability 15, no. 3 (2023): 2204. http://dx.doi.org/10.3390/su15032204.
Full textAbstract:
Because of the rise in the number of cyberattacks, the devices that make up the Internet of Things (IoT) environment are experiencing increased levels of security risks. In recent years, a significant number of centralized systems have been developed to identify intrusions into the IoT environment. However, due to diverse requirements of IoT devices such as dispersion, scalability, resource restrictions, and decreased latency, these strategies were unable to achieve notable outcomes. The present paper introduces two novel metaheuristic optimization algorithms for optimizing the weights of deep learning (DL) models, use of DL may help in the detection and prevention of cyberattacks of this nature. Furthermore, two hybrid DL classifiers, i.e., convolutional neural network (CNN) + deep belief network (DBN) and bidirectional long short-term memory (Bi-LSTM) + gated recurrent network (GRU), were designed and tuned using the already proposed optimization algorithms, which results in ads to improved model accuracy. The results are evaluated against the recent approaches in the relevant field along with the hybrid DL classifier. Model performance metrics such as accuracy, rand index, f-measure, and MCC are used to draw conclusions about the model’s validity by employing two distinct datasets. Regarding all performance metrics, the proposed approach outperforms both conventional and cutting-edge methods.
Dissertations / Theses on the topic "Hybrid deep learning (DL) model"
1
Awan, Ammar Ahmad. "Co-designing Communication Middleware and Deep Learning Frameworks for High-Performance DNN Training on HPC Systems." The Ohio State University, 2020. http://rave.ohiolink.edu/etdc/view?acc_num=osu1587433770960088.
Full text
2
Torregrosa, jordan Sergio. "Approches Hybrides et Méthodes d'Intelligence Artificielle Basées sur la Simulation Numérique pour l'Optimisation des Systèmes Aérodynamiques Complexes." Electronic Thesis or Diss., Paris, HESAM, 2024. http://www.theses.fr/2024HESAE002.
Full textAbstract:
La conception industrielle d'un composant est un processus complexe, long et coûteux, contraint par des spécifications physiques, stylistiques et de développement précises en fonction de ses conditions et de son environnement d'utilisation futurs. En effet, un composant industriel est défini et caractérisé par de nombreux paramètres qui doivent être optimisés pour satisfaire au mieux toutes ces spécifications. Cependant, la complexité de ce problème d'optimisation multiparamétrique sous contraintes est telle que sa résolution analytique est compromise.Dans le passé, un tel problème était résolu expérimentalement par essais et erreurs, entraînant des processus de conception coûteux et chronophages. Depuis le milieu du 20e siècle, avec l'accès généralisé à des moyens de calcul de plus en plus puissants, les ``jumeaux virtuels'' ou simulations numériques basées sur la physique, sont devenus un outil essentiel pour la recherche, réduisant le besoin de mesures expérimentales. À la fin du XXe siècle, le volume de données augmente et se répands massivement dans la plupart des domaines. Ceci conduit à la prolifération des techniques d'Intelligence Artificielle (IA), ou ``jumeaux numériques'', remplaçant partiellement les ``jumeaux virtuels'' grâce à leur plus faible technicité. Aujourd'hui, ces évolutions ont abouti à un cadre où la théorie, l'expérimentation, la simulation et les données peuvent interagir en synergie et se renforcer mutuellement.Dans ce contexte, Stellantis vise à explorer comment l'IA peut améliorer le processus de conception d'un système complexe. A cette fin, l'objectif principal de cette thèse est de développer un modèle de substitution paramétrique de la géométrie d'un aérateur innovant. Le modèle renvoit la norme du champ de vitesse au niveau du visage du pilote afin d'explorer l'espace des géométries possibles tout en évaluant leurs performances en temps réel. Le développement d'un tel modèle basé sur des données pose plusieurs problèmes conceptuels qui peuvent être résolus par l'IA.L'utilisation de techniques de régression classiques peut conduire à des résultats non physiques dans certains domaines tels que la dynamique des fluides. Ainsi, le modèle de substitution paramétrique proposé est basé sur la théorie du Transport Optimal (OT) qui offre une approche mathématique pour mesurer des distances et interpoler d'une manière novatrice.Le succès d'un modèle basé sur des données dépend de la qualité des données d'entraînement. D'une part, les données expérimentales sont considérées comme les plus réalistes, mais elles sont extrêmement coûteuses et laborieuses. D'autre part, les simulations numériques sont plus accessibles et rapides, mais présentent un écart important par rapport à la réalité. Ainsi, une approche Jumeau Hybride est proposée, basée sur la théorie du OT, afin de combler l'ignorance entre la simulation et la mesure.Le processus d'échantillonnage des données d'entraînement est devenu une charge de travail centrale dans le processus de développement d'un modèle basé sur des données. Une méthodologie d'Apprentissage Actif est donc proposée pour sélectionner de manière itérative et intelligente les points d'entraînement, baséee sur les objectifs industriels attendus du composant étudié, afin de minimiser le nombre d'échantillons nécessaires. Ainsi, cette stratégie d'échantillonnage maximise les performances du modèle tout en convergeant vers la solution optimale du problème industriel.L'exactitude d'un modèle basé sur des données est généralement l’objectif principal lors de son entraînement. Or, la réalité est complexe et imprévisible, ce qui fait que des paramètres d'entrée peuvent être connus avec un certain degré d'incertitude. Par conséquent, une méthodologie de quantification des incertitudes, basée sur les estimateurs de Monte Carlo et l'OT, est proposée pour prendre en compte la propagation des incertitudes dans le modèle et pour quantifier leur impact sur sa précision<br>The industrial design of a component is a complex, time-consuming and costly process constrained to precise physical, styling and development specifications led by its future conditions and environment of use. Indeed, an industrial component is defined and characterized by many parameters which must be optimized to best satisfy all those specifications. However, the complexity of this multi-parametric constrained optimization problem is such that its analytical resolution is compromised.In the recent past, such a problem was solved experimentally, by trial and error, leading to expensive and time-consuming design processes. Since the mid-20th century, with the advancement and widespread access to increasingly powerful computing technologies, the ``virtual twins'', or physics-based numerical simulations, became an essential tool for research and development, significantly diminishing the need for experimental measurements. However, despite the computing power available today, ``virtual twins'' are still limited by the complexity of the problem solved and present some significant deviations from reality due to the ignorance of certain subjacent physics. In the late 20th century, the volume of data has surge enormously, massively spreading in the majority of fields and leading to a wide proliferation of Artificial Intelligence (AI) techniques, or ``digital twins'', partially substituting the ``virtual twins'' thanks to their lower intricacy. Nevertheless, they need an important training stage and can lead to some aversion since they operate as black boxes. Today, these technological evolutions have resulted in a framework where theory, experimentation, simulation and data can interact in synergy and reinforce each other.In this context, Stellantis aims to explore how AI can improve the design process of a complex aerodynamic system: an innovative cockpit air vent. To this purpose, the main goal of this thesis is to develop a parametric surrogate of the aerator geometry which outputs the norm of the velocity field at the pilot's face in order to explore the space of possible geometries while evaluating their performances in real time. The development of such a data-based metamodel entails several conceptual problems which can be addressed with AI.The use of classical regression techniques can lead to unphysical interpolation results in some domains such as fluid dynamics. Thus, the proposed parametric surrogate is based on Optimal Transport (OT) theory which offers a mathematical approach to measure distances and interpolate between general objects in a novel way.The success of a data-driven model relies on the quality of the training data. On the one hand, experimental data is considered as the most realistic but is extremely costly and time-consuming. On the other hand, numerical simulations are cheaper and faster but present a significant deviation from reality. Therefore, a Hybrid Twin approach is proposed based on Optimal Transport theory in order to bridge the ignorance gap between simulation and measurement.The sampling process of training data has become a central workload in the development process of a data-based model. Hence, an Active Learning methodology is proposed to iteratively and smartly select the training points, based on industrial objectives expected from the studied component, in order to minimize the number of needed samples. Thus, this sampling strategy maximizes the performance of the model while converging to the optimal solution of the industrial problem.The accuracy of a data-based model is usually the main concern of its training process. However, reality is complex and unpredictable leading to input parameters known with a certain degree of uncertainty. Therefore, a data-based Uncertainty Quantifcation methodology, based on Monte Carlo estimators and OT, is proposed to take into account the uncertainties propagation into the surrogate and to quantify their impact on its precision
3
Budaraju, Sri Datta. "Unsupervised 3D Human Pose Estimation." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-291435.
Full textAbstract:
The thesis proposes an unsupervised representation learning method to predict 3D human pose from a 2D skeleton via a VAEGAN (Variational Autoencoder Generative Adversarial Network) hybrid network. The method learns to lift poses from 2D to 3D using selfsupervision and adversarial learning techniques. The method does not use images, heatmaps, 3D pose annotations, paired/unpaired 2Dto3D skeletons, 3D priors, synthetic 2D skeletons, multiview or temporal information in any shape or form. The 2D skeleton input is taken by a VAE that encodes it in a latent space and then decodes that latent representation to a 3D pose. The 3D pose is then reprojected to 2D for a constrained, selfsupervised optimization using the input 2D pose. Parallelly, the 3D pose is also randomly rotated and reprojected to 2D to generate a ’novel’ 2D view for unconstrained adversarial optimization using a discriminator network. The combination of the optimizations of the original and the novel 2D views of the predicted 3D pose results in a ’realistic’ 3D pose generation. The thesis shows that the encoding and decoding process of the VAE addresses the major challenge of erroneous and incomplete skeletons from 2D detection networks as inputs and that the variance of the VAE can be altered to get various plausible 3D poses for a given 2D input. Additionally, the latent representation could be used for crossmodal training and many downstream applications. The results on Human3.6M datasets outperform previous unsupervised approaches with less model complexity while addressing more hurdles in scaling the task to the real world.<br>Uppsatsen föreslår en oövervakad metod för representationslärande för att förutsäga en 3Dpose från ett 2D skelett med hjälp av ett VAE GAN (Variationellt Autoenkodande Generativt Adversariellt Nätverk) hybrid neuralt nätverk. Metoden lär sig att utvidga poser från 2D till 3D genom att använda självövervakning och adversariella inlärningstekniker. Metoden använder sig vare sig av bilder, värmekartor, 3D poseannotationer, parade/oparade 2D till 3D skelett, a priori information i 3D, syntetiska 2Dskelett, flera vyer, eller tidsinformation. 2Dskelettindata tas från ett VAE som kodar det i en latent rymd och sedan avkodar den latenta representationen till en 3Dpose. 3D posen är sedan återprojicerad till 2D för att genomgå begränsad, självövervakad optimering med hjälp av den tvådimensionella posen. Parallellt roteras dessutom 3Dposen slumpmässigt och återprojiceras till 2D för att generera en ny 2D vy för obegränsad adversariell optimering med hjälp av ett diskriminatornätverk. Kombinationen av optimeringarna av den ursprungliga och den nya 2Dvyn av den förutsagda 3Dposen resulterar i en realistisk 3Dposegenerering. Resultaten i uppsatsen visar att kodningsoch avkodningsprocessen av VAE adresserar utmaningen med felaktiga och ofullständiga skelett från 2D detekteringsnätverk som indata och att variansen av VAE kan modifieras för att få flera troliga 3D poser för givna 2D indata. Dessutom kan den latenta representationen användas för crossmodal träning och flera nedströmsapplikationer. Resultaten på datamängder från Human3.6M är bättre än tidigare oövervakade metoder med mindre modellkomplexitet samtidigt som de adresserar flera hinder för att skala upp uppgiften till verkliga tillämpningar.
4
Bergkvist, Alexander, Nils Hedberg, Sebastian Rollino, and Markus Sagen. "Surmize: An Online NLP System for Close-Domain Question-Answering and Summarization." Thesis, Uppsala universitet, Institutionen för informationsteknologi, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-412247.
Full textAbstract:
The amount of data available and consumed by people globally is growing. To reduce mental fatigue and increase the general ability to gain insight into complex texts or documents, we have developed an application to aid in this task. The application allows users to upload documents and ask domain-specific questions about them using our web application. A summarized version of each document is presented to the user, which could further facilitate their understanding of the document and guide them towards what types of questions could be relevant to ask. Our application allows users flexibility with the types of documents that can be processed, it is publicly available, stores no user data, and uses state-of-the-art models for its summaries and answers. The result is an application that yields near human-level intuition for answering questions in certain isolated cases, such as Wikipedia and news articles, as well as some scientific texts. The application shows a decrease in reliability and its prediction as to the complexity of the subject, the number of words in the document, and grammatical inconsistency in the questions increases. These are all aspects that can be improved further if used in production.<br>Mängden data som är tillgänglig och konsumeras av människor växer globalt. För att minska den mentala trötthet och öka den allmänna förmågan att få insikt i komplexa, massiva texter eller dokument, har vi utvecklat en applikation för att bistå i de uppgifterna. Applikationen tillåter användare att ladda upp dokument och fråga kontextspecifika frågor via vår webbapplikation. En sammanfattad version av varje dokument presenteras till användaren, vilket kan ytterligare förenkla förståelsen av ett dokument och vägleda dem mot vad som kan vara relevanta frågor att ställa. Vår applikation ger användare möjligheten att behandla olika typer av dokument, är tillgänglig för alla, sparar ingen personlig data, och använder de senaste modellerna inom språkbehandling för dess sammanfattningar och svar. Resultatet är en applikation som når en nära mänsklig intuition för vissa domäner och frågor, som exempelvis Wikipedia- och nyhetsartiklar, samt viss vetensaplig text. Noterade undantag för tillämpningen härrör från ämnets komplexitet, grammatiska korrekthet för frågorna och dokumentets längd. Dessa är områden som kan förbättras ytterligare om den används i produktionen.
5
Mohy, El Dine Kamal. "Control of robotic mobile manipulators : application to civil engineering." Thesis, Université Clermont Auvergne (2017-2020), 2019. http://www.theses.fr/2019CLFAC015/document.
Full textAbstract:
Malgré le progrès de l'automatisation industrielle, les solutions robotiques ne sont pas encore couramment utilisées dans le secteur du génie civil. Plus spécifiquement, les tâches de ponçage, telles que le désamiantage, sont toujours effectuées par des opérateurs humains utilisant des outils électriques et hydrauliques classiques. Cependant, avec la diminution du coût relatif des machines par rapport au travail humain et les réglementations sanitaires strictes applicables à des travaux aussi risqués, les robots deviennent progressivement des alternatives crédibles pour automatiser ces tâches et remplacer les humains.Dans cette thèse, des nouvelles approches de contrôle de ponçage de surface sont élaborées. Le premier contrôleur est un contrôleur hybride position-force avec poignet conforme. Il est composé de 3 boucles de commande, force, position et admittance. La commutation entre les commandes pourrait créer des discontinuités, ce qui a été résolu en proposant une commande de transition. Dans ce contrôleur, la force de choc est réduite par la commande de transition proposée entre les modes espace libre et contact. Le second contrôleur est basé sur un modèle de ponçage développé et un contrôleur hybride adaptatif position-vitesse-force. Les contrôleurs sont validés expérimentalement sur un bras robotique à 7 degrés de liberté équipé d'une caméra et d'un capteur de force-couple. Les résultats expérimentaux montrent de bonnes performances et les contrôleurs sont prometteurs. De plus, une nouvelle approche pour contrôler la stabilité des manipulateurs mobiles en temps réel est présentée. Le contrôleur est basé sur le « zero moment point », il a été testé dans des simulations et il a été capable de maintenir activement la stabilité de basculement du manipulateur mobile tout en se déplaçant. En outre, les incertitudes liées à la modélisation et aux capteurs sont prises en compte dans les contrôleurs mentionnés où des observateurs sont proposés.Les détails du développement et de l'évaluation des différents contrôleurs proposés sont présentés, leurs mérites et leurs limites sont discutés et des travaux futurs sont suggérés<br>Despite the advancements in industrial automation, robotic solutions are not yet commonly used in the civil engineering sector. More specifically, grinding tasks such as asbestos removal, are still performed by human operators using conventional electrical and hydraulic tools. However, with the decrease in the relative cost of machinery with respect to human labor and with the strict health regulations on such risky jobs, robots are progressively becoming credible alternatives to automate these tasks and replace humans.In this thesis, novel surface grinding control approaches are elaborated. The first controller is based on hybrid position-force controller with compliant wrist and a smooth switching strategy. In this controller, the impact force is reduced by the proposed smooth switching between free space and contact modes. The second controller is based on a developed grinding model and an adaptive hybrid position-velocity-force controller. The controllers are validated experimentally on a 7-degrees-of-freedom robotic arm equipped with a camera and a force-torque sensor. The experimental results show good performances and the controllers are promising. Additionally, a new approach for controlling the stability of mobile manipulators in real time is presented. The controller is based on zero moment point, it is tested in simulations and it was able to actively maintain the tip-over stability of the mobile manipulator while moving. Moreover, the modeling and sensors uncertainties are taken into account in the mentioned controllers where observers are proposed. The details of the development and evaluation of the several proposed controllers are presented, their merits and limitations are discussed and future works are suggested
6
WEI, JING-LING, and 魏敬玲. "A Hybrid Model using Deep Learning for Crowding Status Prediction at Emergency Departments." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/448qrt.
Full textAbstract:
碩士<br>國立中正大學<br>資訊工程研究所<br>106<br>Over the last two decades, the number of patients at emergency departments in Taiwan has grown significantly. According to the statistics, in a year, nearly 220,000 patients visit the emergency department of the hospital which has the largest number of visits. An average of 600 patients pour into the emergency department each day. The increasing number of emergency department visits and the emergency department crowding have become major public health problems globally. The ability to accurately predict the level of demand (i.e., predict patient flow) has considerable implications for hospitals’ development of good management strategies and resource allocation. The core objective of this research is to predict the crowding status (i.e., free, moderately crowded, severely crowded, and overloaded) of an emergency department for the next time period. Meanwhile, we can predict the operational states of the emergency department including the actual number of patients and beds for the next time period so as to help the emergency department administrators to optimize planning. In this research, we combine deep learning with conventional machine learning techniques to construct a three-stage hybrid prediction model which utilizes the SVM classification algorithm and LSTM for prediction. In the first stage, an individual LSTM model for each feature is constructed to predict the value of each feature for the next 30 minutes. In the second stage, we use SVM to establish the rules of each crowding status of the hospitals. According to these rules, we can understand the correlation between the crowding status of the emergency department and the circumstances of a given hospital. We use the newly predicted records that were generated by the LSTM models for the rules that were established by the SVM model to predict the crowding status for the next time period.
7
EVANDER, RICHARD, and RICHARD EVANDER. "A Hybrid Deep Machine Learning Model For Soil Classification of Compressed Earth Block." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/9x3u39.
Full textAbstract:
碩士<br>國立臺灣科技大學<br>營建工程系<br>107<br>Classifying or predicting soil type for Compressed Earth Block (CEB) construction using machine learning model is an important technique to replace laboratory tests which are time and cost consuming. The previous study has established soil classification using Artificial Neural Network (ANN). Nonetheless, gradient-based learning on ANN face several issues like overfitting and trapped in local minima due to poor generalization performance. In the further development of the neural network model, Extreme Learning Machine (ELM) has been developed with faster learning speed and better generalization performance compared to ANN. However, ELM itself has some drawbacks in random input weights and heavy memory problems. This research proposed the model called Backpropagated Deep Restricted Boltzmann Extreme Learning Machine (BD-RBELM) which establish deep ELM network to solve the heavy memory problem and use Restricted Boltzmann Machine (RBM) to train input weight for solving random input weights problem. The model then used for the soil classification problem with the final performance of 95.144% accuracy and 95.120 F1-Score, which performed better compared with the other AI techniques.
8
Huang, Shen-Hang, and 黃慎航. "Online Structural Break Detection for Pairs Trading using Wavelet Transform and Hybrid Deep Learning Model." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/5kkcxr.
Full textAbstract:
碩士<br>國立交通大學<br>資訊科學與工程研究所<br>108<br>With the mature development in the financial market, numerous people study in arbitrage strategies. Pairs trading is one of the common statistical arbitrage strategies. It first supervises two stocks that move similarly and form a stationary equilibrium with certain weights, and then makes arbitrage when the pair deviates from the stable value. The time point that the stationary relationship between two stocks does not exist any longer is called a structural break, and detecting structural breaks is important to pairs trading. There are some traditional methods for this problem, but they are not robust enough to implement in the real world. The purpose of this paper is to precisely detect structural breaks as soon as possible. Therefore, we propose a hybrid wavelet transform deep learning model using both frequency-domain and time-domain features to detect a structural break of a stock pair in Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX). We collect the amount of half-year historical tick data for experiments and build a simulation trading system to evaluate the performances of traditional methods and our models in the real condition. The experiment results on performance metrics and simulation trading show that our proposed method successfully not only captures the abnormal signal but also reduces the loss occurred from structural breaks.
9
HSIEH, YI-LIN, and 謝易霖. "High Dimensional Deep Learning of Real-Time Stock Price Forecasting Model by Hybrid Dimension Reduction Method." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/c9y4ec.
Full textAbstract:
碩士<br>輔仁大學<br>統計資訊學系應用統計碩士班<br>106<br>Nowadays in Taiwan people find themselves hard to pay living expenses just by their salaries, and stocks became a popular choice to gain wealth. Stock Price varies with many unpredictable messages or some unperceivable complicated relations, so there are many variables to consider about. If there are ways good enough to reduce dimensions and get features that really changes stock price, it will be able to determine trends and get more remuneration.
So, this research uses real time information of Taiwanese stock market, western and some Asian index along with information of Institutional investors related with Margin Trading and Short Selling in Taiwan Stock Exchange, and use autoencoder to reduce dimension and predict the stock prices of three targets: MediaTek, Getac and CTBC Financial Holding. However, the dimension reduces by autoencoder didn't take the effects by response variable into consideration. So we submitted a resolution by adding random forest and change point detection to select further outcomes made by autoencoder.
Our results proved that compared to using all variables or use only autoencoder, using autoencoder to decrease dimensions and use random forest and change point detection can get lower RMSE.
10
Partovi, Tahmineh. "3D Building Model Reconstruction from Very High Resolution Satellite Stereo Imagery." Doctoral thesis, 2019. https://repositorium.ub.uni-osnabrueck.de/handle/urn:nbn:de:gbv:700-201910022067.
Full textAbstract:
Automatic three-dimensional (3D) building model reconstruction using remote sensing data is crucial in applications which require large-scale and frequent building model updates, such as disaster monitoring and urban management, to avoid huge manual efforts and costs. Recent advances in the availability of very high-resolution satellite data together with efficient data acquisition and large area coverage have led to an upward trend in their applications for 3D building model reconstructions. In this dissertation, a novel multistage hybrid automatic 3D building model reconstruction approach is proposed which reconstructs building models in level of details 2 (LOD2) based on digital surface model (DSM) data generated from the very high-resolution stereo imagery of the WorldView-2 satellite. This approach uses DSM data in combination with orthorectified panchromatic (PAN) and pan-sharpened data of multispectral satellite imagery to overcome the drawbacks of DSM data, such as blurred building boundaries, rough building shapes unwanted failures in the roof geometries. In the first stage, the rough building boundaries in the DSM-based building masks are refined by classifying the geometrical features of the corresponding PAN images. The refined boundaries are then simplified in the second stage through a parameterization procedure which represents the boundaries by a set of line segments. The main orientations of buildings are then determined, and the line segments are regularized accordingly. The regularized line segments are then connected to each other based on a rule-based method to form polygonal building boundaries. In the third stage, a novel technique is proposed to decompose the building polygons into a number of rectangles under the assumption that buildings are usually composed of rectangular structures. In the fourth stage, a roof model library is defined, which includes flat, gable, half-hip, hip, pyramid and mansard roofs. These primitive roof types are then assigned to the rectangles based on a deep learning-based classification method. In the fifth stage, a novel approach is developed to reconstruct watertight parameterized 3D building models based on the results of the previous stages and normalized DSM (nDSM) of satellite imagery. In the final stage, a novel approach is proposed to optimize building parameters based on an exhaustive search, so that the two-dimensional (2D) distance between the 3D building models and the building boundaries (obtained from building masks and PAN image) as well as the 3D normal distance between the 3D building models and the 3D point clouds (obtained from nDSM) are minimized. Different parts of the building blocks are then merged through a newly proposed intersection and merging process. All corresponding experiments were conducted on four areas of the city of Munich including 208 buildings and the results were evaluated qualitatively and quantitatively. According to the results, the proposed approach could accurately reconstruct 3D models of buildings, even the complex ones with several inner yards and multiple orientations. Furthermore, the proposed approach provided a high level of automation by the limited number of primitive roof model types required and by performing automatic parameter initialization. In addition, the proposed boundary refinement method improved the DSM-based building masks specified by 8 % in area accuracy. Furthermore, the ridge line directions and roof types were detected accurately for most of the buildings. The combination of the first three stages improved the accuracy of the building boundaries by 70 % in comparison to using line segments extracted from building masks without refinement. Moreover, the proposed optimization approach could achieve in most cases the best combinations of 2D and 3D geometrical parameters of roof models. Finally, the intersection and merging process could successfully merge different parts of the complex building models.
Book chapters on the topic "Hybrid deep learning (DL) model"
1
Camargo, Manuel, Marlon Dumas, and Oscar González-Rojas. "Learning Accurate Business Process Simulation Models from Event Logs via Automated Process Discovery and Deep Learning." In Advanced Information Systems Engineering. Springer International Publishing, 2022. http://dx.doi.org/10.1007/978-3-031-07472-1_4.
Full textAbstract:
AbstractBusiness process simulation is a well-known approach to estimate the impact of changes to a process with respect to time and cost measures – a practice known as what-if process analysis. The usefulness of such estimations hinges on the accuracy of the underlying simulation model. Data-Driven Simulation (DDS) methods leverage process mining techniques to learn process simulation models from event logs. Empirical studies have shown that, while DDS models adequately capture the observed sequences of activities and their frequencies, they fail to accurately capture the temporal dynamics of real-life processes. In contrast, generative Deep Learning (DL) models are better able to capture such temporal dynamics. The drawback of DL models is that users cannot alter them for what-if analysis due to their black-box nature. This paper presents a hybrid approach to learn process simulation models from event logs wherein a (stochastic) process model is extracted via DDS techniques, and then combined with a DL model to generate timestamped event sequences. An experimental evaluation shows that the resulting hybrid simulation models match the temporal accuracy of pure DL models, while partially retaining the what-if analysis capability of DDS approaches.
2
Chukwu, Emmanuel C., and Pedro A. Moreno-Sánchez. "Enhancing Arrhythmia Diagnosis with Data-Driven Methods: A 12-Lead ECG-Based Explainable AI Model." In Communications in Computer and Information Science. Springer Nature Switzerland, 2024. http://dx.doi.org/10.1007/978-3-031-59091-7_16.
Full textAbstract:
AbstractAccurate and early prediction of arrhythmias using Electrocardiograms (ECG) presents significant challenges due to the non-stationary nature of ECG signals and inter-patient variability, posing difficulties even for seasoned cardiologists. Deep Learning (DL) methods offer precision in identifying diagnostic ECG patterns for arrhythmias, yet they often lack the transparency needed for clinical application, thus hindering their broader adoption in healthcare. This study introduces an explainable DL-based prediction model using ECG signals to classify nine distinct arrhythmia categories. We evaluated various DL architectures, including ResNet, DenseNet, and VGG16, using raw ECG data. The ResNet34 model emerged as the most effective, achieving an Area Under the Receiver Operating Characteristic (AUROC) of 0.98 and an F1-score of 0.826. Additionally, we explored a hybrid approach that combines raw ECG signals with Heart Rate Variability (HRV) features. Our explainability analysis, utilizing the SHAP technique, identifies the most influential ECG leads for each arrhythmia type and pinpoints critical signal segments for individual disease prediction. This study emphasizes the importance of explainability in arrhythmia prediction models, a critical aspect often overlooked in current research, and highlights its potential to enhance model acceptance and utility in clinical settings.
3
Soderi, Simone, Mariella Särestöniemi, Syifaul Fuada, Matti Hämäläinen, Marcos Katz, and Jari Iinatti. "Securing Hybrid Wireless Body Area Networks (HyWBAN): Advancements in Semantic Communications and Jamming Techniques." In Communications in Computer and Information Science. Springer Nature Switzerland, 2024. http://dx.doi.org/10.1007/978-3-031-59091-7_24.
Full textAbstract:
AbstractThis paper explores novel strategies to strengthen the security of Hybrid Wireless Body Area Networks (HyWBANs), which are essential in smart healthcare and Internet of Things (IoT) applications. Recognizing the vulnerability of HyWBAN to sophisticated cyber-attacks, we propose an innovative combination of semantic communications and jamming receivers. This dual-layered security mechanism protects against unauthorized access and data breaches, particularly in scenarios involving in-body to on-body communication channels. We conduct comprehensive laboratory measurements to understand hybrid (radio and optical) communication propagation through biological tissues. We utilize these insights to refine a dataset for training a Deep Learning (DL) model. These models, in turn, generate semantic concepts linked to cryptographic keys for enhanced data confidentiality and integrity using a jamming receiver. The proposed model significantly reduces energy consumption compared to traditional cryptographic methods, like Elliptic Curve Diffie-Hellman (ECDH), especially when supplemented with jamming. Our approach addresses the primary security concerns and sets the baseline for future secure biomedical communication systems advancements.
4
Rippel, Oliver, and Dorit Merhof. "Anomaly Detection for Automated Visual Inspection: A Review." In Bildverarbeitung in der Automation. Springer Berlin Heidelberg, 2023. http://dx.doi.org/10.1007/978-3-662-66769-9_1.
Full textAbstract:
AbstractAnomaly detection (AD) methods that are based on deep learning (DL) have considerably improved the state of the art in AD performance on natural images recently. Combined with the public release of large-scale datasets that target AD for automated visual inspection (AVI), this has triggered the development of numerous, novel AD methods specific to AVI. However, with the rapid emergence of novel methods, the need to systematically categorize them arises. In this review, we perform such a categorization, and identify the underlying assumptions as well as working principles of DL-based AD methods that are geared towards AVI. We perform this for 2D AVI setups, and find that the majority of successful AD methods currently combines features generated by pre-training DL models on large-scale, natural image datasets with classical AD methods in hybrid AD schemes. Moreover, we give the main advantages and drawbacks of the two identified model categories in the context of AVI’s inherent requirements. Last, we outline open research questions, such as the need for an improved detection performance of semantic anomalies, and propose potential ways to address them.
5
Verma, Jyoti, Isha Kansal, Renu Popli, et al. "A Hybrid Images Deep Trained Feature Extraction and Ensemble Learning Models for Classification of Multi Disease in Fundus Images." In Communications in Computer and Information Science. Springer Nature Switzerland, 2024. http://dx.doi.org/10.1007/978-3-031-59091-7_14.
Full textAbstract:
AbstractRetinal disorders, including diabetic retinopathy and macular degeneration due to aging, can lead to preventable blindness in diabetics. Vision loss caused by diseases that affect the retinal fundus cannot be reversed if not diagnosed and treated on time. This paper employs deep-learned feature extraction with ensemble learning models to improve the multi-disease classification of fundus images. This research presents a novel approach to the multi-classification of fundus images, utilizing deep-learned feature extraction techniques and ensemble learning to diagnose retinal disorders and diagnosing eye illnesses involving feature extraction, classification, and preprocessing of fundus images. The study involves analysis of deep learning and implementation of image processing. The ensemble learning classifiers have used retinal photos to increase the classification accuracy. The results demonstrate improved accuracy in diagnosing retinal disorders using DL feature extraction and ensemble learning models. The study achieved an overall accuracy of 87.2%, which is a significant improvement over the previous study. The deep learning models utilized in the study, including NASNetMobile, InceptionResNetV4, VGG16, and Xception, were effective in extracting relevant features from the Fundus images. The average F1-score for Extra Tree was 99%, while for Histogram Gradient Boosting and Random Forest, it was 98.8% and 98.4%, respectively. The results show that all three algorithms are suitable for the classification task. The combination of DenseNet feature extraction technique and RF, ET, and HG classifiers outperforms other techniques and classifiers. This indicates that using DenseNet for feature extraction can effectively enhance the performance of classifiers in the task of image classification.
6
Panner Selvam, Karthick, and Mats Brorsson. "DIPPM: A Deep Learning Inference Performance Predictive Model Using Graph Neural Networks." In Euro-Par 2023: Parallel Processing. Springer Nature Switzerland, 2023. http://dx.doi.org/10.1007/978-3-031-39698-4_1.
Full textAbstract:
AbstractDeep Learning (DL) has developed to become a corner-stone in many everyday applications that we are now relying on. However, making sure that the DL model uses the underlying hardware efficiently takes a lot of effort. Knowledge about inference characteristics can help to find the right match so that enough resources are given to the model, but not too much. We have developed a DL Inference Performance Predictive Model (DIPPM) that predicts the inference latency, energy, and memory usage of a given input DL model on the NVIDIA A100 GPU. We also devised an algorithm to suggest the appropriate A100 Multi-Instance GPU profile from the output of DIPPM. We developed a methodology to convert DL models expressed in multiple frameworks to a generalized graph structure that is used in DIPPM. It means DIPPM can parse input DL models from various frameworks. Our DIPPM can be used not only helps to find suitable hardware configurations but also helps to perform rapid design-space exploration for the inference performance of a model. We constructed a graph multi-regression dataset consisting of 10,508 different DL models to train and evaluate the performance of DIPPM, and reached a resulting Mean Absolute Percentage Error (MAPE) as low as 1.9%.
7
Bhandari, Palak, Chetan R. Patil, Chetan S. Patil, Shivtej Deshmukh, and Ranjana Badre. "Resume Screening Using Hybrid Deep Learning Model." In Multi-Strategy Learning Environment. Springer Nature Singapore, 2024. http://dx.doi.org/10.1007/978-981-97-1488-9_43.
Full text
8
Khatchadourian, Raffi, Tatiana Castro Vélez, Mehdi Bagherzadeh, Nan Jia, and Anita Raja. "Hybridize Functions: A Tool for Automatically Refactoring Imperative Deep Learning Programs to Graph Execution." In Lecture Notes in Computer Science. Springer Nature Switzerland, 2025. https://doi.org/10.1007/978-3-031-90900-9_5.
Full textAbstract:
Abstract Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code—supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. Though hybrid approaches aim for the “best of both worlds,” using them effectively requires subtle considerations to make code amenable to safe, accurate, and efficient graph execution—avoiding performance bottlenecks and semantically inequivalent results. We discuss the engineering aspects of a refactoring tool that automatically determines when it is safe and potentially advantageous to migrate imperative DL code to graph execution and vice-versa.
9
Disabato, Simone. "Deep and Wide Tiny Machine Learning." In Special Topics in Information Technology. Springer International Publishing, 2022. http://dx.doi.org/10.1007/978-3-031-15374-7_7.
Full textAbstract:
AbstractIn the last decades, on the one hand, Deep Learning (DL) has become state of the art in several domains, e.g., image classification, object detection, and natural language processing. On the other hand, pervasive technologies—Internet of Things (IoT) units, embedded systems, and Micro-Controller Units (MCUs)—ask for intelligent processing mechanisms as close as possible to data generation. Nevertheless, memory, computational, and energy requirements characterizing DL models are three or more orders of magnitude larger than the corresponding memory, computation, and energy capabilities of pervasive devices. This work aims at introducing a methodology to address this issue and enable pervasive intelligent processing. In particular, by defining Tiny Machine Learning (TML) solutions, i.e., machine and deep learning models that take into account the constraints on memory, computation, and energy of the target pervasive device. The proposed methodology addresses the problem at three different levels. In the first approach, the methodology devices inference-based Deep TML solutions by approximation techniques, i.e., the TML model runs on the pervasive device but was trained elsewhere. Then, the methodology introduces on-device learning for TML. Finally, the third approach develops Wide Deep TML solutions that split and distribute the DL processing over connected heterogeneous pervasive devices.
10
Berkani, Lamia, Dyhia Laga, and Abdelhak Aissat. "Social Neural Hybrid Recommendation with Deep Representation Learning." In Model and Data Engineering. Springer International Publishing, 2021. http://dx.doi.org/10.1007/978-3-030-78428-7_11.
Full textConference papers on the topic "Hybrid deep learning (DL) model"
1
Hancı, Nur Banu, İlke Kurt, Sezer Ulukaya, Oğuzhan Erdem, Sibel Güler, and Cem Uzun. "Hybrid Voice Spectrogram-Chromogram Based Deep Learning (HVSC-DL) Model for the Detection of Parkinson’s Disease." In 2024 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA). IEEE, 2024. http://dx.doi.org/10.23919/spa61993.2024.10715598.
Full text
2
Chhaglani, Bhawana, Ebrahim Nemati, Sharath Chandrashekhara, Jilong Kuang, and Alex Gao. "Cough-DL: A Deep Learning Model for Ear-Worn Cough Detection." In 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). IEEE, 2024. https://doi.org/10.1109/embc53108.2024.10782702.
Full text
3
Chaturvedi, Rajnish Kumar, and Nitin Arvind Shelke. "Music generation using hybrid deep neural model." In 2024 Intelligent Systems and Machine Learning Conference (ISML). IEEE, 2024. https://doi.org/10.1109/isml60050.2024.11007294.
Full text
4
Chandira Prabha, S., S. Kaviyadharshini, and A. Annie Micheal. "A Hybrid Deep Learning Model for Violence Detection." In 2025 3rd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT). IEEE, 2025. https://doi.org/10.1109/idciot64235.2025.10914952.
Full text
5
Win, Aye Mya Mya, and Ah Nge Htwe. "Abnormal Fall Detection by Hybrid Deep Learning Model." In 2025 IEEE Conference on Computer Applications (ICCA). IEEE, 2025. https://doi.org/10.1109/icca65395.2025.11011186.
Full text
6
M, Bhavani, and Prithi Samuel. "Automated Hematocyte Classification using Hybrid Deep Learning Model." In 2025 5th International Conference on Pervasive Computing and Social Networking (ICPCSN). IEEE, 2025. https://doi.org/10.1109/icpcsn65854.2025.11035474.
Full text
7
Prakash, Ravi, Trilok Nath Pandey, Bibhuti Bhusan Dash, Sudhansu Shekhar Patra, Utpal Chandra De, and Abinash Tripathy. "Skin Cancer Diagnosis using Deep Learning, Transfer Learning and Hybrid Model." In 2024 Second International Conference on Inventive Computing and Informatics (ICICI). IEEE, 2024. http://dx.doi.org/10.1109/icici62254.2024.00024.
Full text
8
BalaKrishna, N., R. Sai Vikas Reddy, S. Geethika, M. Likhith, N. Lasya Priya, and S. H. Harsha. "Tackling Depression Detection with deep learning a Hybrid Model." In 2024 International Conference on Knowledge Engineering and Communication Systems (ICKECS). IEEE, 2024. http://dx.doi.org/10.1109/ickecs61492.2024.10617226.
Full text
9
Aanand, B. Akshai, Frederick Ruby Helen, and B. Lavaraju. "A Deep Learning-Based Hybrid Model for Enhancing Security." In 2025 International Conference on Visual Analytics and Data Visualization (ICVADV). IEEE, 2025. https://doi.org/10.1109/icvadv63329.2025.10961476.
Full text
10
Angalees, Biju, and Anupama D. Sakhare. "Efficient Marine Fish Classification Using Hybrid Deep Learning Model." In 2024 IEEE 2nd International Conference on Innovations in High Speed Communication and Signal Processing (IHCSP). IEEE, 2024. https://doi.org/10.1109/ihcsp63227.2024.10959748.
Full textReports on the topic "Hybrid deep learning (DL) model"
1
Ogunbire, Abimbola, Panick Kalambay, Hardik Gajera, and Srinivas Pulugurtha. Deep Learning, Machine Learning, or Statistical Models for Weather-related Crash Severity Prediction. Mineta Transportation Institute, 2023. http://dx.doi.org/10.31979/mti.2023.2320.
Full textAbstract:
Nearly 5,000 people are killed and more than 418,000 are injured in weather-related traffic incidents each year. Assessments of the effectiveness of statistical models applied to crash severity prediction compared to machine learning (ML) and deep learning techniques (DL) help researchers and practitioners know what models are most effective under specific conditions. Given the class imbalance in crash data, the synthetic minority over-sampling technique for nominal (SMOTE-N) data was employed to generate synthetic samples for the minority class. The ordered logit model (OLM) and the ordered probit model (OPM) were evaluated as statistical models, while random forest (RF) and XGBoost were evaluated as ML models. For DL, multi-layer perceptron (MLP) and TabNet were evaluated. The performance of these models varied across severity levels, with property damage only (PDO) predictions performing the best and severe injury predictions performing the worst. The TabNet model performed best in predicting severe injury and PDO crashes, while RF was the most effective in predicting moderate injury crashes. However, all models struggled with severe injury classification, indicating the potential need for model refinement and exploration of other techniques. Hence, the choice of model depends on the specific application and the relative costs of false negatives and false positives. This conclusion underscores the need for further research in this area to improve the prediction accuracy of severe and moderate injury incidents, ultimately improving available data that can be used to increase road safety.
2
Pasupuleti, Murali Krishna. Decision Theory and Model-Based AI: Probabilistic Learning, Inference, and Explainability. National Education Services, 2025. https://doi.org/10.62311/nesx/rriv525.
Full textAbstract:
Abstract Decision theory and model-based AI provide the foundation for probabilistic learning, optimal inference, and explainable decision-making, enabling AI systems to reason under uncertainty, optimize long-term outcomes, and provide interpretable predictions. This research explores Bayesian inference, probabilistic graphical models, reinforcement learning (RL), and causal inference, analyzing their role in AI-driven decision systems across various domains, including healthcare, finance, robotics, and autonomous systems. The study contrasts model-based and model-free approaches in decision-making, emphasizing the trade-offs between sample efficiency, generalization, and computational complexity. Special attention is given to uncertainty quantification, explainability techniques, and ethical AI, ensuring AI models remain transparent, accountable, and risk-aware. By integrating probabilistic reasoning, deep learning, and structured decision models, this research highlights how AI can make reliable, interpretable, and human-aligned decisions in complex, high-stakes environments. The findings underscore the importance of hybrid AI frameworks, explainable probabilistic models, and uncertainty-aware optimization, shaping the future of trustworthy, scalable, and ethically responsible AI-driven decision-making. Keywords Decision theory, model-based AI, probabilistic learning, Bayesian inference, probabilistic graphical models, reinforcement learning, Markov decision processes, uncertainty quantification, explainable AI, causal inference, model-free learning, Monte Carlo methods, variational inference, hybrid AI frameworks, ethical AI, risk-aware decision-making, optimal control, trust in AI, interpretable machine learning, autonomous systems, financial AI, healthcare AI, AI governance, explainability techniques, real-world AI applications.
3
Pasupuleti, Murali Krishna. Optimal Control and Reinforcement Learning: Theory, Algorithms, and Robotics Applications. National Education Services, 2025. https://doi.org/10.62311/nesx/rriv225.
Full textAbstract:
Abstract: Optimal control and reinforcement learning (RL) are foundational techniques for intelligent decision-making in robotics, automation, and AI-driven control systems. This research explores the theoretical principles, computational algorithms, and real-world applications of optimal control and reinforcement learning, emphasizing their convergence for scalable and adaptive robotic automation. Key topics include dynamic programming, Hamilton-Jacobi-Bellman (HJB) equations, policy optimization, model-based RL, actor-critic methods, and deep RL architectures. The study also examines trajectory optimization, model predictive control (MPC), Lyapunov stability, and hierarchical RL for ensuring safe and robust control in complex environments. Through case studies in self-driving vehicles, autonomous drones, robotic manipulation, healthcare robotics, and multi-agent systems, this research highlights the trade-offs between model-based and model-free approaches, as well as the challenges of scalability, sample efficiency, hardware acceleration, and ethical AI deployment. The findings underscore the importance of hybrid RL-control frameworks, real-world RL training, and policy optimization techniques in advancing robotic intelligence and autonomous decision-making. Keywords: Optimal control, reinforcement learning, model-based RL, model-free RL, dynamic programming, policy optimization, Hamilton-Jacobi-Bellman equations, actor-critic methods, deep reinforcement learning, trajectory optimization, model predictive control, Lyapunov stability, hierarchical RL, multi-agent RL, robotics, self-driving cars, autonomous drones, robotic manipulation, AI-driven automation, safety in RL, hardware acceleration, sample efficiency, hybrid RL-control frameworks, scalable AI.