5916 research outputs found
Sort by
A single-stage step-down rectifier (S3R) for AC-DC power conversion
No full textThe actively controlled AC/DC power conversion has a limitation in certain moments that the output voltage could be higher than twice the peak value of the supply voltage. A Single-Stage Step-Down Rectifier ((SR)-R-3) is proposed to overcome this limitation by having a half-bridge rectifier and a buck (step-down) converter in single-stage. Both the half-bridge rectifier and the DC/DC step-down converter require two switches each, resulting in a total of four switches in the system. However, the proposed rectifier reduces the number of switches to three by connecting them in series within a single-leg configuration. The output inductor (step-down inductor) operates in continuous conduction mode (CCM) with a designed pulse width modulation (PWM) strategy. The proposed rectifier has a computed 2.83% total harmonic distortion (THD) and achieves almost unity power factor (PF) for 9.9 kW output power at 240 V-rms line voltage, in the simulation. Simulation results performed in MATLAB SIMULINK confirmed the feasibility and functionality of the proposed single-stage topology
Resistance under confinement: resilience of protests and their limits in authoritarian Turkey
No full textIn this paper, we examine the relationship between the process of autocratisation and protests, and argue that scholarship on electoral autocracies should not only focus on major protest cycles but also examine 'ordinary' protests to understand how social and political actors resist and push back against autocratisation. Using an original dataset of protest events from 2015 to 2021, we analyse the transformation of protests in Turkey as it experienced gradual but significant autocratisation. We discuss two mechanisms through which autocratisation might affect levels, actors and repertoires of protesting: first, via increasing repression; and, second, via the policy choices of the authoritarian regime. Our findings indicate that protests continued even under the state of emergency in Turkey, but with significant changes in levels and repertoires of protesting. The protest scene was dominated by protests using tactics that rely on a small number of individuals and are contained in their spatial reach and disruptiveness. This research underlines the importance of examining ordinary protests to analyse how autocratisation transforms protests, using original data from local sources.Bogazici Universit
The effect of the emdr flash group technique on test anxiety, traumatic stress, and coping mechanisms in students taking a gap year
No full textN/APublisher versio
The role of sugar sources and locality claims on front-of-package labels: Effects on perceptions and purchase intentions among university students in Türkiye
No full textThis study examines the effects of front-of-package (FOP) claims related to sugar source and locality on university students’ perceptions and purchase intentions in Türkiye. An experimental design was employed with four packaging conditions: (1) a control package with no claim, (2) a package labeled “Made with Beet Sugar,” (3) one labeled “Taste of Anatolia,” and (4) a package displaying both claims. Participants (n = 104) viewed the packaging, tasted cookies, and completed a questionnaire assessing product liking, purchase intention, health motivation, perceived healthiness, and perceived quality. Results revealed no significant differences in perceived healthiness or product liking across packaging conditions. However, participants exposed to the dual-claim condition reported significantly lower purchase intentions compared to the control group. Regression analyses showed that health motivation significantly predicted both perceived healthiness and purchase intention. Moreover, product liking, perceived quality, and perceived healthiness jointly predicted purchase intention, with product liking being the most influential factor. These findings underscore the limited impact of sugar source and locality claims on consumer evaluations in this context and highlight the role of individual traits, such as health motivation, in shaping responses to FOP labels.TÜBİTA
Practical applications of AI tools in foreign language education: Opportunities and challenges
No full textThe integration of Artificial Intelligence (AI) into education has gained increasing attention in recent years, offering teachers and learners new possibilities to enhance learning processes. This paper explores the practical applications of AI tools in foreign language education with a focus on their pedagogical potential, advantages, and limitations. The discussion categorizes AI tools into four main groups: text-based, visual, audio-video, and translation technologies. Text-based tools such as ChatGPT, QuillBot, and Rytr support writing activities by assisting learners with brainstorming, paraphrasing, and editing, thereby encouraging critical thinking and creativity. Visual tools including Canva, DALL-E, and Midjourney promote creativity and communication by enabling the design of posters, infographics, and illustrations that enrich language learning tasks. Audio and video tools such as Murf.ai, Synthesia, and Adobe Podcast provide authentic speaking and listening opportunities, enhance learner motivation, and help teachers prepare multimodal materials efficiently. Translation tools such as DeepL, Google Lens, and Microsoft Translator facilitate multilingual communication, making foreign language learning more accessible and inclusive. Findings from the literature and classroom practices indicate that these tools contribute significantly to the development of 21st-century skills, particularly creativity, communication, collaboration, and critical thinking. However, several challenges remain, including ethical concerns, issues of data privacy, overreliance on automated outputs, limitations in contextual accuracy, and unequal access to technology across regions. The paper argues that AI should not be regarded as a replacement for teachers but rather as a complementary support that enhances pedagogical practices when integrated thoughtfully. By highlighting both the opportunities and challenges of AI in foreign language education, this study emphasizes the importance of equipping teachers with digital literacy and ethical awareness to ensure responsible and effective use of these technologies
Financial statement fraud detection with a categorical-to-numerical data representation
No full textIdentifying fraudulent financial reports and elucidating the mechanisms of fraud are critical for safeguarding investors from substantial losses. Financial statements present detailed accounting entries in tabular form; they inherently combine categorical and numerical variables governed by accounting dependencies, yet most existing methods fail to model interpretable interactions between these feature types. In this case, handling categorical variables together with numerical variables is important in enhancing the financial statement fraud detection performance. Here, we compare the methods for transforming categorical to numerical attributes, which are then used for financial statement fraud detection. We perform comprehensive experiments on two real-world datasets: FiGraph and USFSD. We compare 4 state-of-the-art specialized categorical-to-numerical transformation techniques with several other simpler statistical encoding mechanisms, such as target, label, Helmert, and GLMM encodings, as well as methods that can directly work on categorical data, such as CatBoost. These specialized transformation techniques are Hierarchical Coupling Learning-based CURE, Graph-based Categorical Embedding GCE, and Transitive Distance Learning-based embedding. The results reveal that the performance of CURE and XGBoost together surpasses all state-of-the-art techniques, achieving significant relative gains in macro-level recall over the second-best performing approaches, CatBoost and FTTransformer, while also providing clear and interpretable insights into the discovered fraud pathways.TÜBİTAKPublisher versio
Laser-induced breakdown spectroscopy by nJ pulses at GHz repetition rate
No full textFiber lasers operating at GHz pulse repetition rate have shown exceptional performance in high-speed high-efficient material processing [1]. In laser-induced breakdown spectroscopy (LIBS), fiber lasers have attracted significant interest because of their advantages, including compactness, energy efficiency, stability, reliability, and cost-effectiveness. To the best of our knowledge, we employed for the first time a burst mode with GHz intraburst repetition rate pulses in laser-induced breakdown spectroscopy, reducing the required pulse energy by several orders of magnitude to the nJ level. We employed a custom-built Yb-doped 1040 nm fiber laser operating in burst mode with an intraburst repetition rate of 2.8 GHz and pulse energies ranging from 9 nJ to 196 nJ [2]. The system generates 40 ps pulses before compression and 650 fs pulses after compression, with a peak average power 14 W. LIBS experiments were carried out on stainless steel (SS) at a burst repetition rate of 100 kHz, using three different burst durations: 83 ns, 120 ns and 240 ns. We investigate the influences of burst duration and pulse energy on the optical emission spectrum of SS. As depicted in Fig.(a) it was observed that the blackbody-like spectrum which appeared while using high burst duration 240 ns, due to significant thermal effects, as most of the pulse energy is absorbed as heat by the target, is reduced when using the shorter burst durations. At GHz pulse repetition rate, this reduction in the ablation threshold enables materials to be ablated at much lower energy levels, generating a highly ionized plasma with intense emissions. As a result the thermal effects are minimized, leading to a spectrum with sharp, well-defined peaks. Subsequently, the impact of the laser pulse energy on the spectrum of SS was examined as displayed in Fig(b), at lower energies (26-57 nJ) the emission lines were relatively weak, and the signal to noise ratio (SNR) is also low. This was attributed to insufficient material ablation, as energy is near to ablation threshold, with most of it being used for plasma production. As the laser energy increases (57-196 nJ) the intensity of the emission lines increases significantly due to more effective material ablation and enhanced plasma excitation. Moreover, the SNR improves and the continuum background intensifies as the plasma temperature as the plasma temperature increases at higher energies, causing bremsstrahlung radiations to become dominant. In the second part of this study we focused on calculating the electron temperature and density to investigate how they vary with different burst durations and laser pulse energies. The Boltzmann method was used to calculate the electron temperature Te, while the Stark broadened profile of the 537.14 Fe I line was analyzed to estimate the electron density ne. At a burst energy 45.5 µJ the electron temperature was measured as 1812 K, 1600 K and 1514 K for burst duration 83 ns, 120 ns and 240 ns respectively. Similarly, the corresponding electron densities were found to be 1.959×1017, 1.874×1017, and 1.386×1017cm-3. It was concluded that the electron temperature and density rise with a decrease in the burst durations as shown in Fig.(c)&(d)
Artificial pain representation with tactile and vision blending
No full textAs robots become increasingly embedded in human environments, the ability to anticipate the outcomes of physical contact is crucial for enabling safe, adaptive, and socially intelligent behavior. Thus, learning to discriminate harmful sensory patterns from the benign ones will not only ensure physical safety during robot interaction, but may also lay the foundation for artificial empathy through mirroring the pain of others. To this end, this work develops a framework for tactile prediction through multimodal learning, emphasizing the integration of visual and tactile information in a common latent space. The ability to predict tactile sensations prior to contact allows a robot to avoid harmful outcomes as well as internalizing the tactile experience of others. We adapt the Deep Modality Blending Network (DMBN) as a foundational model for this task. Using demonstrations involving both gentle and noxious human touch, synchronized visual and tactile data are collected to train the model. After learning, the robot can generate temporal tactile activations from visual observations alone, anticipating sensory outcomes before physical contact occurs. Experiments on an upper-body humanoid robot show that it can predict painful stimuli and mirror tactile experiences observed in others. The key contributions of this study include: (1) the development of a predictive tactile perception framework using DMBNs, (2) the adaptation of this framework for modeling artificial pain that may be used as a basis for artificial empathy, and (3) empirical validation using real-world humanrobot interaction scenarios. © 2025 IEEE.LARSyS FCT ; Japan Society for the Promotion of Scienc
Effect of strain rate on the mechanical behavior of additivemanufactured aluminum alloy after severe plastic deformation
No full textThis work presents the mechanical behavior of additive-manufactured AlSiMg alloy after severe plastic deformation (SPD). Equal Channel Angular Extrusion/Pressing (ECAE/P) is a wellknown SPD method used for mechanical property improvement via grain refinement in metals. In this study, 8 pass ECAP at 250°C ECAP process is conducted on AlSi10Mg which is manufactured by laser powder bed fusion (LPBF). Tensile tests were conducted at room temperature and at various strain rates to measure the strain rate sensitivity (SRS). With varying strain rate, there was appreciable change in the flow stress levels indicating that the severely deformed alloy exhibits negative SRS. Possible reasons for this mechanical response are explained based on the evolved microstructure to shed light on the parameters governing SRS in additive-manufactured alloys subjected to SPD.TÜBİTA
CS-REG-NET: Termal ve optik görüntülerde çapraz-spektral çakıştırma için görsel-durum uzayı tabanlı özgözetimli ögrenmeli mimari
No full textModern deep models for multispectral image matching typically rely on large, supervised datasets, which can be prohibitively expensive. To overcome this challenge, we introduce CS-REG-NET, a self-supervised, detector-based framework that requires no external labels. Instead, it uses RIFT2 detector to generate pseudo-ground-truth keypoints. A VMamba encoder, pre-trained on a segmentation task, processes image pairs, while two output heads learn feature heatmaps and descriptors. CSREG-NET significantly outperforms existing methods, delivering superior keypoint detection and homography estimation. This real-time framework thus provides a robust, extensible solution for multispectral image matching.TÜBİTA