10 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 10 / 10
Yayın Computer aided darlington synthesis of an all purpose immittance function(Istanbul University, 2016) Yarman, Bekir Sıddık Binboğa; Aksen, Ahmet; Köprü, Ramazan; Kumar, Narendra Senthil; Aydın, Çağatay; Atilla, Doğu Çağdaş; Chacko, PrakashThis work is the continuation of our high precision immittance synthesis paper series introduced in IEEE TCAS-I. In the present manuscript, we modified the previously introduced high precision Bandpass LC-ladder synthesis algorithm to include the extraction of finite frequency and right half plane (RHP) transmission zeros of an impedance function as Brune/Darlington Type-C sections. Finite frequency and RHP transmission zeros are extracted employing our newly introduced modified impedance and chain parameters based algorithms one by one. After each transmission zero extraction, remaining immittance function is corrected using parametric approach. It is shown that propsed high precision synthesis algorithms can synthesize immittance functions up to 40 reactive elements with accumulated relative error in the order of 10- 1 . The modified high precision synthesis package is developed in MatLab environment and it is integrated with the real frequency techniques to design matching networks over broadbands. Examples are presented to exhibit the usage of the newly proposed high precision synthesis algorithms.Yayın Vikipedi ve Vikisözlük'ten Hypernym çıkarma(IEEE, 2017-06-27) Şaşmaz, Emre; Ehsani, Razieh; Yıldız, Olcay TanerDoğal dil işleme alanında kullanılan önemli yapılardan bir tanesi WordNet gibi büyük ölçekli sözlüklerdir. WordNet; eşanlamlı, zıt anlamlı gibi anlamsal ilişkileri de içeren kapsamlı bir sözlüktür. Bu bildiride, WordNet’in önemli bir parçası olan Hypernym-Hyponym ilişkisini çıkarmaya çalıştık. Bu amaca ulaşmak için, Vikipedi, Türkçe Sözlük ve Vikisözlük kaynaklarını kullandık. Sonlu Durum Makinelerinden ürettiğimiz kurallarla Hypernym-Hyponym ilişkilerini çıkardık.Yayın The composition of acids in bitumen and in products from saponification of kerogen: Investigation of their role as connecting kerogen and mineral matrix(Elsevier Science BV, 2008-11-03) Razvigorova, Maria; Budinova, Temenuzhka K.; Tsyntsarski, Boyko G.; Petrova, Bilyana N.; Ekinci, Ekrem; Atakül, HüsnüIn order to obtain more information and to understand the nature of relation between organic and mineral matter in oil shales, the compositions of soluble bitumen fractions obtained by extraction from Bulgarian oil shales before and after demineralization with 10% HCl, concentrated HE and a HF/HCl mixture were investigated. The four extracts were quantitatively examined by IR and H-1 NMR spectroscopy. The investigation of isolated acidic material of the bitumen fractions showed that the fatty acids are present in bitumen fractions as free acids, esters and salts. The amount of free acids in bitumen is very small. The dominant part of bitumen acids is associated with mineral components of the oil shales as well as part of them is included in the mineral matrix, and can be separated only after deep demineralization. The kerogen of the oil shales, obtained after separation of the bitumen fractions and mineral components, was subjected to saponification in order to determine the amount of acids, bound as esters to the kerogen matrix. The major components found were n-carboxylic, alpha,omega,-di-carboxylic, and aromatic acids. The connection of kerogen with mineral components is accomplished by the participation of carboxylic and complicated ester bonds. Experimental data for the composition of bitumen acids give evidence that algae and terrestrial materials are initial sources in the formation of soluble organic matter of Bulgarian oil shale.Yayın Unsupervised textile defect detection using convolutional neural networks(Elsevier Ltd, 2021-12) Koulali, Imane; Eskil, Mustafa TanerIn this study, we propose a novel motif-based approach for unsupervised textile anomaly detection that combines the benefits of traditional convolutional neural networks with those of an unsupervised learning paradigm. It consists of five main steps: preprocessing, automatic pattern period extraction, patch extraction, features selection and anomaly detection. This proposed approach uses a new dynamic and heuristic method for feature selection which avoids the drawbacks of initialization of the number of filters (neurons) and their weights, and those of the backpropagation mechanism such as the vanishing gradients, which are common practice in the state-of-the-art methods. The design and training of the network are performed in a dynamic and input domain-based manner and, thus, no ad-hoc configurations are required. Before building the model, only the number of layers and the stride are defined. We do not initialize the weights randomly nor do we define the filter size or number of filters as conventionally done in CNN-based approaches. This reduces effort and time spent on hyper-parameter initialization and fine-tuning. Only one defect-free sample is required for training and no further labeled data is needed. The trained network is then used to detect anomalies on defective fabric samples. We demonstrate the effectiveness of our approach on the Patterned Fabrics benchmark dataset. Our algorithm yields reliable and competitive results (on recall, precision, accuracy and f1-measure) compared to state-of-the-art unsupervised approaches, in less time, with efficient training in a single epoch and a lower computational cost.Yayın Work and heat value of bound entanglement(Springer, 2019-12) Tuncer, Aslı; Izadyari, Mohsen; Müstecaplıoğlu, Özgür Esat; Özaydın, Fatih; Daǧ, Ceren B.Entanglement has recently been recognized as an energy resource which can outperform classical resources if decoherence is relatively low. Multi-atom entangled states can mutate irreversibly to so-called bound entangled (BE) states under noise. Resource value of BE states in information applications has been under critical study, and a few cases where they can be useful have been identified. We explore the energetic value of typical BE states. Maximal work extraction is determined in terms of ergotropy. Since the BE states are nonthermal, extracting heat from them is less obvious. We compare single and repeated interaction schemes to operationally define and harvest heat from BE states. BE and free entangled (FE) states are compared in terms of their ergotropy and maximal heat values. Distinct roles of distillability in work and heat values of FE and BE states are pointed out. Decoherence effects in dynamics of ergotropy and mutation of FE states into BE states are examined to clarify significance of the work value of BE states. Thermometry of distillability of entanglement using micromaser cavity is proposed.Yayın An observation based muscle model for simulation of facial expressions(Elsevier Science BV, 2018-05) Erkoç, Tuğba; Ağdoğan, Didem; Eskil, Mustafa TanerThis study presents a novel facial muscle model for coding of facial expressions. We derive this model from unintrusive observation of human subjects in the progress of the surprise expression. We use a generic and single-layered face model which embeds major muscles of the human face. This model is customized onto the human subject's face on the first frame of the video. The last frame of the video is used to project a set of manually marked feature points to estimate the 3 dimensional displacements of vertices due to facial expression. Vertex displacements are used in a mass spring model to estimate the external forces, i.e. the muscle forces on the skin. We observed that the distribution of muscle forces resemble sigmoid or hyperbolic tangent functions. We chose hyperbolic tangent function as our base model and parameterized it using least squares. We compared the proposed muscle model with frequently used models in the literature.Yayın Computer aided high precision Darlington synthesis for real frequency matching(Institute of Electrical and Electronics Engineers Inc, 2014) Yarman, Bekir Sıddık Binboğa; Aksen, Ahmet; Köprü, Ramazan; Aydın, Çağatay; Atilla, Doğu ÇağdaşIn this work, we introduce a high precision synthesis algorithm to include the extraction of finite frequency and right half plane (RHP) transmission zeros of an impedance function as Brune/Darlington Type-C sections. After each transmission zero extraction, remaining immittance function is corrected using a parametric approach. It is shown that proposed high precision synthesis algorithm can synthesize immittance functions up to 40 reactive elements with accumulated relative error in the order of 10-1. The high precision synthesis package is integrated with the real frequency techniques to design matching networks over broadbands. Examples are presented to exhibit the usage of the proposed high precision synthesis algorithm.Yayın Tunneling rates of single electrons on liquid helium in an extracting field(Springer/Plenum Publishers, 2009-02) Karakurt, İsmailWe calculated the tunneling rates of single electrons from the quasi-stationary states on both liquid He-4 and He-3 in an extracting electric field. The rates were obtained from the widths of the resonance lineshapes of the asymptotic amplitude of the wave function for the electron. The calculations were carried out in the limit of strong tunneling which leads to tunneling rates of the order of 1 GHz, and is of recent interest in a proposed quantum-computer system using the electronic states of electrons on helium as qubits. We find that the resonances can, in general, be described by Fano lineshapes. Our results, in addition to presenting quantitative information involving the read-out operations of qubits, clarify that when tunneling is weak the resonances are sharp and more accurately Lorentzian and that the observed Fano lineshapes result from strong tunneling leading to fat resonances and hence asymmetry.Yayın A novel similarity based unsupervised technique for training convolutional filters(IEEE, 2023-05-17) Erkoç, Tuğba; Eskil, Mustata TanerAchieving satisfactory results with Convolutional Neural Networks (CNNs) depends on how effectively the filters are trained. Conventionally, an appropriate number of filters is carefully selected, the filters are initialized with a proper initialization method and trained with backpropagation over several epochs. This training scheme requires a large labeled dataset, which is costly and time-consuming to obtain. In this study, we propose an unsupervised approach that extracts convolutional filters from a given dataset in a self-organized manner by processing the training set only once without using backpropagation training. The proposed method allows for the extraction of filters from a given dataset in the absence of labels. In contrast to previous studies, we no longer need to select the best number of filters and a suitable filter weight initialization scheme. Applying this method to the MNIST, EMNIST-Digits, Kuzushiji-MNIST, and Fashion-MNIST datasets yields high test performances of 99.19%, 99.39%, 95.03%, and 90.11%, respectively, without applying backpropagation training or using any preprocessed and augmented data.Yayın BOUN-ISIK participation: an unsupervised approach for the named entity normalization and relation extraction of Bacteria Biotopes(Association for Computational Linguistics (ACL), 2019-11-04) Karadeniz, İlknur; Tuna, Ömer Faruk; Özgu, ArzucanThis paper presents our participation at the Bacteria Biotope Task of the BioNLP Shared Task 2019. Our participation includes two systems for the two subtasks of the Bacteria Biotope Task: the normalization of entities (BB-norm) and the identification of the relations between the entities given a biomedical text (BB-rel). For the normalization of entities, we utilized word embeddings and syntactic re-ranking. For the relation extraction task, pre-defined rules are used. Although both approaches are unsupervised, in the sense that they do not need any labeled data, they achieved promising results. Especially, for the BB-norm task, the results have shown that the proposed method performs as good as deep learning based methods, which require labeled data.
