Search

Your search keyword '"Keskin AS"' showing total 543 results
Start Over Author "Keskin AS" Publisher ieee
543 results on '"Keskin AS"'

5. Yapay Öğrenme ile El Yazılarından Kişilik Özellikleri Çıkarımı

Author

Keskin, M. Mert, Özbayoğlu, A. Murat, Yüksel, N. İdil, Keskin, M. Mert, Özbayoğlu, A. Murat, and Yüksel, N. İdil

Abstract

Grafoloji çalışma alanında insanların el yazılarıyla kişilik özellikleri arasında bağlantı kurulmaya çalışılır. Bu çalışmada yapay öğrenme ve psikoloji bilimi kullanılarak grafolojinin bilimselliği üzerine bir analiz yapılmıştır. Çalışmaya gönüllü ˘ olarak katılan kullanıcılara bir temel kişilik testi uygulanmış, aynı kullanıcılardan alınan yazı örnekleri üzerinden bazı ayırt edici yazı karakteristikleri çıkarılmış, ve bu yazı karakteristikleri üzerinden oluşturulan özniteliklere bağlı yapay öğrenme tabanlı ˘ tahmin modelleri geliştirilmiştir. Burada amaç kullanıcının yazı stilinden otomatik olarak kişilik analizi ile ilgili ne kadar başarılı bir eşleştirme yapılabileceğinin incelenmesidir. Elde edilen sonuçlar yapay öğrenme modellerinin yazı stilinden kişilik özelliklerini ˘ çok hassas olmasa da belirli bir ölçüde tahmin edebildiğini göstermektedir., In the field of graphology, the fundamental idea is to establish a connection between the handwriting style of people and their personality traits. In this study, an analysis was made on the science of graphology using machine learning and psychology. A basic personality test was applied to the users who voluntarily participated in our study, some distinctive writing characteristics were extracted from the text samples taken from the same users, and machine learning-based prediction models were developed based on the features constructed over these font characteristics. Here, the aim is to examine how successful matching can be implemented automatically from the user’s writing style with regard to personality analysis. The results show that even though not very sensitive, the personality traits can be predicted to a certain extent from the writing style by the machine learning models.

Published
2024

6. An Optimum Volumetric Array Design Approach for Both Azimuth and Elevation Isotropic DOA Estimation

Author

Fesih Keskin and Tansu Filik

Subjects
2-D isotropic array, 3-D arrays, angular ambiguity, array signal processing, direction of arrival, DOA, Electrical engineering. Electronics. Nuclear engineering, TK1-9971
Abstract

In this study, a new volumetric array design approach is presented for both azimuth and elevation isotropic direction of arrival (DOA) estimation. The presented method uses a minimum number of additional sensors to extend the given arrays to a volumetric array which has equal values for the Cramér-Rao lower bound (CRLB) for the entire azimuth and elevation angles. This approach can be applied to all arbitrary linear and planar arrays. Moreover, the design approach also takes into account the electromagnetic mutual coupling (MC) effect between the array elements and both the azimuth and elevation angular ambiguity uncertainties in an optimum manner while extending the array. It has been verified both analytically and with simulations that the designed volumetric arrays satisfy the two dimensional (2-D) isotropic DOA estimation conditions. The proposed approach is applied to Uniform Linear Arrays (ULA), Uniform Circular Arrays (UCA), V-shaped Arrays, and an arbitrary planar array to extend to optimum volumetric arrays with considerable results.

Published
2020
Full Text
View/download PDF

8. A Comparison of Loss Weighting Strategies for Multi task Learning in Deep Neural Networks

Author

Ting Gong, Tyler Lee, Cory Stephenson, Venkata Renduchintala, Suchismita Padhy, Anthony Ndirango, Gokce Keskin, and Oguz H. Elibol

Subjects
Dynamic weighting average, multi-MNIST, multi-objective optimization, multi-task learning, uncertainty weighting, Electrical engineering. Electronics. Nuclear engineering, TK1-9971
Abstract

With the success of deep learning in a wide variety of areas, many deep multi-task learning (MTL) models have been proposed claiming improvements in performance obtained by sharing the learned structure across several related tasks. However, the dynamics of multi-task learning in deep neural networks is still not well understood at either the theoretical or experimental level. In particular, the usefulness of different task pairs is not known a priori. Practically, this means that properly combining the losses of different tasks becomes a critical issue in multi-task learning, as different methods may yield different results. In this paper, we benchmarked different multi-task learning approaches using shared trunk with task specific branches architecture across three different MTL datasets. For the first dataset, i.e. Multi-MNIST (Modified National Institute of Standards and Technology database), we thoroughly tested several weighting strategies, including simply adding task-specific cost functions together, dynamic weight average (DWA) and uncertainty weighting methods each with various amounts of training data per-task. We find that multi-task learning typically does not improve performance for a user-defined combination of tasks. Further experiments evaluated on diverse tasks and network architectures on various datasets suggested that multi-task learning requires careful selection of both task pairs and weighting strategies to equal or exceed the performance of single task learning.

Published
2019
Full Text
View/download PDF

19. Monostatic Sensing with OFDM under Phase Noise

Author

Keskin, Musa Furkan, Wymeersch, Henk, Koivunen, Visa, Chalmers University of Technology, Visa Koivunen Group, Department of Information and Communications Engineering, Aalto-yliopisto, and Aalto University

Subjects
Radar tracking, Radar, iterated small angle approximation, Sensors, joint radar-communications, Phase locked loops, Delays, Estimation, exploitation, phase noise, OFDM
Abstract

Publisher Copyright: IEEE We consider the problem of monostatic radar sensing with orthogonal frequency-division multiplexing (OFDM) joint radar-communications (JRC) systems in the presence of phase noise (PN) caused by oscillator imperfections. We begin by providing a rigorous statistical characterization of PN in the radar receiver over multiple OFDM symbols for free-running oscillators (FROs) and phase-locked loops (PLLs). Based on the delay-dependent PN covariance matrix, we derive the hybrid maximum-likelihood (ML)/maximum a-posteriori (MAP) estimator of the deterministic delay-Doppler parameters and the random PN, resulting in a challenging high-dimensional nonlinear optimization problem. To circumvent the nonlinearity of PN, we then develop an iterated small angle approximation (ISAA) algorithm that progressively refines delay-Doppler-PN estimates via closed-form updates of PN as a function of delay-Doppler at each iteration. Moreover, unlike existing approaches where PN is considered to be purely an impairment that has to be mitigated, we propose to exploit PN for resolving range ambiguity by capitalizing on its delay-dependent statistics (i.e., the range correlation effect), through the formulation of a parametric Toeplitz-block Toeplitz covariance matrix reconstruction problem. Simulation results indicate quick convergence of ISAA to the hybrid Cramér-Rao bound (CRB), as well as its remarkable performance gains over state-of-the-art benchmarks, for both FROs and PLLs under various operating conditions, while showing that the detrimental effect of PN can be turned into an advantage for sensing.

Published
2023

20. Rethinking the Data Annotation Process for Multiview 3D Pose Estimation with Active Learning and Self-Training

Author

Feng, Qi, He, Kun, Wen, He, Keskin, Cem, and Ye, Yuting

Subjects
FOS: Computer and information sciences, Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition
Abstract

Pose estimation of the human body and hands is a fundamental problem in computer vision, and learning-based solutions require a large amount of annotated data. In this work, we improve the efficiency of the data annotation process for 3D pose estimation problems with Active Learning (AL) in a multi-view setting. AL selects examples with the highest value to annotate under limited annotation budgets (time and cost), but choosing the selection strategy is often nontrivial. We present a framework to efficiently extend existing single-view AL strategies. We then propose two novel AL strategies that make full use of multi-view geometry. Moreover, we demonstrate additional performance gains by incorporating pseudo-labels computed during the AL process, which is a form of self-training. Our system significantly outperforms simulated annotation baselines in 3D body and hand pose estimation on two large-scale benchmarks: CMU Panoptic Studio and InterHand2.6M. Notably, on CMU Panoptic Studio, we are able to reduce the turn-around time by 60% and annotation cost by 80% when compared to the conventional annotation process., IEEE WACV 2023 algorithms track. Code: https://github.com/facebookresearch/multi_view_active_learning

Published
2023
Full Text
View/download PDF

38. 6G Radio Requirements to Support Integrated Communication, Localization, and Sensing

Author

Wymeersch, Henk, primary, Parssinen, Aarno, additional, Abrudan, Traian E., additional, Wolfgang, Andreas, additional, Haneda, Katsuyuki, additional, Sarajlic, Muris, additional, Leinonen, Marko E., additional, Keskin, Musa Furkan, additional, Chen, Hui, additional, Lindberg, Simon, additional, Kyosti, Pekka, additional, Svensson, Tommy, additional, and Yang, Xinxin, additional

Published
2022
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results