5916 research outputs found
Sort by
Vispesar: Video summarization by prompt engineering and serverless architecture
No full textThis paper presents VISPESAR, a cloud-based video summarization service that extracts short clips from uploaded videos based on natural language queries or prompts. The current implementation uses Google Gemini API to interpret user prompts and identify content that matches user queries. FFmpeg software is used for video processing, mainly for keyframe extraction, video slicing, and concatenation to generate summaries. The backend employs a serverless architecture with Amazon Web Services (AWS) Lambda, Simple Storage Service (S3), Step Functions, and DynamoDB, while providing real-time feedback to the frontend user interfaces through WebSocket communication. The overall solution demonstrates how cloud computing and Artificial Intelligence (AI) can be used together to enable affordable and scalable video intelligence. Through careful design and experimentation, we find that quality of a modern video summarization service depends on: 1-clarity of the prompts used, 2-accuracy of the object detection and classification from frames, and 3- efficiency and scalability of the cloud-based system implementation. © 2025 IEEE
Far-right parties and gender policies in contemporary Türkiye
No full textThis article examines three far-right parties in contemporary T & uuml;rkiye, which gained coalition potential following the 2017 electoral law revision: the New Welfare Party (YRP), the Free Cause Party (H & Uuml;DAPAR) and the Victory Party (ZP). This study explores their ideological alignment, particularly their stances on family values, gender equality, LGBTQ issues and feminism. While YRP and ZP align with the populist radical right through their nativist, populist and authoritarian ideologies, H & Uuml;DAPAR is categorised as extreme right due to its critical stance on democracy. In addition, YRP and H & Uuml;DAPAR promote conservative views on family values, gender equality and LGBTQ rights, whereas ZP adopts a more liberal position, appealing to voters in socioeconomically developed regions. The findings highlight the diverse characteristics of far-right parties in a non-Western context and their potential for comparative analysis
When firm-level connections corrupt the frontline: How political networks drive unethical selling and upselling behavior
No full textIn business-to-business markets, firms' business and political ties are often treated as strategic resources. Yet little is known about how these macro-level networks shape micro-level behavior, particularly that of frontline salespeople. Drawing on role theory, we argue that business and political ties act as implicit scripts that influence salespeople's role perceptions, ethical boundaries, and selling behaviors. Drawing on evidence from three experimental studies and a qualitative investigation with B2B salespeople, we demonstrate that political ties elevate the propensity for unethical selling, subsequently amplifying upselling behavior. Conversely, business ties foster more restrained selling norms, reducing both unethical conduct and upselling. We further demonstrate that these effects are conditioned by salesperson-specific factors: promotion-focused individuals and those who perceive low relationship continuity are especially responsive to politically embedded cues, whereas preventionfocused salespeople and those anticipating long-term relational continuity exhibit greater restraint. We extend role theory by introducing role ethicality-the salesperson's perception of what constitutes ethically appropriate conduct within their role-as a key mechanism linking firm-level affiliations to individual behavior. Our study demonstrates how a company's external relationship strategies can impact important commercial behaviors in ways that may not be obvious or intended by management.Ozyegin Universit
Finsentiment: Predicting financial sentiment through transfer learning
No full textThere is an increasing interest in financial text mining tasks. Significant progress has been made by using deep learning-based models on a generic corpus, which also shows reasonable results on financial text mining tasks such as financial sentiment analysis. However, financial sentiment analysis is still demanding work because of the insufficiency of labeled data for the financial domain and its specialized language. General-purpose deep learning methods are not as effective mainly due to specialized language used in the financial context. In this study, we focus on enhancing the performance of financial text mining tasks by improving the existing pretrained language models via NLP transfer learning. Pretrained language models demand a small quantity of labeled samples, and they could be enhanced to a greater extent by training them on domain-specific corpora instead. We propose an enhanced model FinSentiment, which incorporates enhanced versions of a number of recently proposed pretrained models, such as BERT, XLNet, RoBERTa, GPT, Llama, and T5, to better perform across NLP tasks in financial domain by training these models on financial domain corpora. The corresponding finance-specific models in FinSentiment are called Fin-BERT, Fin-XLNet, Fin-RoBERTa, Fin-GPT, Fin-Llama, and Fin-T5, respectively. We also propose variants of these models jointly trained over financial domain and general corpora. Our finance-specific FinSentiment models, in general, show the best performance across three financial sentiment analysis datasets, even when only a subpart of these models is fine-tuned with a smaller training set. Our results exhibit enhancement for each tested performance criteria on the existing results for these datasets. Extensive experimental results demonstrate the effectiveness and robustness of especially RoBERTa pretrained on financial corpora. Overall, we show that NLP transfer learning techniques are favorable solutions to financial sentiment analysis tasks. Our source code has been deposited at https://github.com/seferlab/finsentiment. © 2025 John Wiley & Sons Ltd.TÜBİTA
Emulating perceptual development in deep reinforcement learning
No full textThe process of learning in infants differs from the traditional reinforcement learning (RL) methods in several aspects. The biggest difference is that RL assumes a stationary world and an agent with fixed sensory and motor abilities. In contrast, infant development proceeds by unfolding new perceptual and motor abilities in parallel to learning. In spite of the general notion that this staged learning leads for faster and better learning in biological systems, it is not clear how such a learning mechanism can be embedded into a reinforcement learning scenario. In this study, towards this direction, we explored how an emulated perceptual development (EPD) in an RL setting can benefit the learning. As a test bed, we took the Pong game and required the RL agent to learn to play against a pre-programmed opponent by using a policy gradient based deep RL method. During learning, inspired by the progressive perceptual development in infants, the state space representation of the RL agent was changed in stages by incorporating additional information about the environment, which largely invalidated the classical RL assumption. By comparing the proposed perceptual development based learning with the performance of baseline learners, we assessed whether the benefits of developmental learning could be transferred to deep reinforcement learning systems. The results obtained suggest that a suitable perceptual development regime may improve the learning progress and yield better performing agents. © 2025 IEEE.University of Osaka ; Japan Science and Technology Agenc
All-in-one femtosecond laser fabrication and welding of multilayer glass microfluidic devices
No full textWe demonstrate an integrated approach for fabricating multilayer glass microfluidic devices using a single femtosecond laser system for both micromachining and welding. Our custom-designed fiber laser system operates at 1030 nm wavelength with 300 fs pulse duration and repetition rate of 400 kHz - 1 MHz, enabling precise channel fabrication followed by direct glass-glass welding. The system employs a unique spiral scanning strategy for welding, achieving uniform energy deposition and robust bonds while preserving microfluidic features. Fabricated three-layer devices show excellent optical transparency with transmission values comparable to unwelded glass, and the welded interfaces demonstrate complete hermetic sealing necessary for microfluidic applications
Learning secondary tool affordances from human actions using the iCub robot
No full textTools and other objects offer agents a range of potential actions, commonly referred to as affordances. Each tool is typically designed with a primary purpose in mind-like a hammer's function to drive nails. However, tools can also serve purposes beyond their original design. These alternative uses represent secondary affordances, extending the tool's utility beyond its primary intended function. While prior robotics research on affordance perception and learning has primarily focused on primary affordances, our work addresses the less-explored area of learning secondary tool affordances from human partners. Using the iCub robot equipped with three cameras, we observed humans performing actions on twenty objects using four different tools in ways that deviate from their primary purposes. For example, the iCub observed humans using rulers not for measuring but to push, pull, and move objects. In this setting, we constructed a dataset by taking pictures of objects before and after each action is executed. To model secondary affordance learning, we trained three neural networks (ResNet-18, ResNet-50, and ResNet-101) on three prediction tasks using these raw images as input: (1) identifying which tool was used to move an object, (2) predicting the tool with additional action category information, and (3) jointly predicting both the tool and action performed. Our results demonstrate that deep learning architectures enable the iCub robot to successfully predict secondary tool affordances, thereby paving the road for human-robot collaborative object manipulation involving complex affordances. Code and data from this study are available at https://github.com/BosongDing/second_affordance
A crisis within crises: Narratives of food insecurity and coping for syrian refugees in Türkiye
No full textSyrian refugees in Türkiye face significant challenges in accessing adequate and nutritious food, a situation exacerbated by their precarious living conditions and recent external shocks like the COVID-19 pandemic, economic crisis, and the 2023 earthquake. Yet, little research has examined how these refugees perceive their food needs and cope with food insecurity. This article fills this gap by utilizing the Food and Agriculture Organization's framework for studying food security, which encompasses availability, access, utilization, and stability. Through 27 semi-structured interviews, the study explores how refugees handle food insecurity, including by limiting food quantity and diversity, borrowing money and food, and prioritizing children's nutrition. Furthermore, the study highlights the governance dimension of food security, demonstrating how policies and market forces affect refugees' vulnerability to hunger. By adopting a qualitative approach focused on migrants' narratives and experiences, this article provides a nuanced, micro-level understanding of food insecurity. It examines the cultural and social aspects of this issue for refugees, while also offering recommendations for policymakers and humanitarian organizations who aim to mitigate this crisis within a crisis.TÜBİTA
Evaluating the association for the accreditation of architectural education (MIAK) in Türkiye: A comparative analysis within the framework of the canberra accord and UNESCO-UIA standards
No full textThis study analyzes the Architectural Accreditation Board of Turkey (MİAK) within both national and international contexts. Using comparative content analysis, MİAK's criteria and procedures were evaluated against those of the UNESCO–UIA Charter and the member organisations of the Canberra Accord. The findings reveal that MİAK has significantly advanced the quality and international recognition of Turkish architectural education by aligning its standards with global frameworks. Strong convergence is observed in core areas such as design and technical competencies, while divergence remains in leadership and professional ethics. The study also identifies ambiguities in how these dimensions are articulated across international systems. By positioning MİAK within global accreditation discourse, the research highlights its contributions to educational quality assurance and professional harmonisation, while underscoring the need for greater emphasis on professional formation and ethical practice. This analysis fills a notable gap in comparative accreditation literature. © 2025 John Wiley & Sons Ltd