Search Results (31)

In an era where artificial intelligence is rapidly expanding into creative domains, the challenge of modeling human-like cognition and emotion in generative processes becomes increasingly central. The present study was made in connection with the exhibition of Munch’s works held in Rome from February to June 2025. Indeed, the paper introduces the concept of a Cognitive Digital Twin grounded in the Super Time-Cognitive Neural Network (STCNN) framework and applies it to the case of Edvard Munch, the iconic Norwegian expressionist. The proposed system—Munch Digital Twin—goes beyond static generative models by integrating temporal, emotional, and cognitive dimensions through a complex-valued time representation t = a + i·b, where a denotes chronological time and b encodes imagination, memory, and creativity. We define Edge Knowledge as an output-stage re-ranking criterion that admits a generated response only where corpus evidence, knowledge-graph constraints and the LLM surface jointly agree (the boundary, or ‘edge’, between documented identity and machine inference). STCNN allows this twin to process real inputs (text, visual prompts, emotional cues) and generate outputs that reflect both the rational and expressive styles of Munch. The imaginary components of the network enable speculative and affective expansions of known artworks—such as reinterpreting The Scream under new emotional or social contexts. This paper presents the theoretical underpinnings of cognitive digital twins, the architecture of the STCNN-based model, and a prototype implementation trained on Munch’s paintings, letters, and critical essays. The system—comprising a GPT-4-Turbo cloud profile and a 4-bit LLaMA-2-13B edge profile for language, Stable Diffusion 1.5 + LoRA for image generation, a Neo4j knowledge graph, and FAISS retrieval—is trained on approximately 600 letters, 100 artworks, and Munch’s diaries and criticism, and evaluated across 100 interactive sessions with 14 students and expert raters. Headline results against an unconditioned baseline include CLIPScore +13.8%, FID −25.5% (small-sample, indicative), and emotion-cosine similarity +44.9%. Ethical implications surrounding posthumous digital emulation, authorship, and emotional manipulation are also discussed. The Munch Digital Twin represents a new paradigm in AI-driven art, where machines do not merely replicate, but collaborate across time with human legacies, enabling an anticipatory and emotionally intelligent form of computational creativity. This work is primarily a conceptual and architectural contribution, supported by a proof-of-concept prototype and a preliminary, non-controlled user study; the quantitative results are indicative and not yet confirmatory. Full article

This study presents the development of an AI-powered chatbot designed to facilitate accurate and efficient retrieval of information from the FDA drug labeling documents. Leveraging OpenAI’s GPT-3.5-turbo model within a controlled, document-grounded question–answering framework, Chatbot was created, which can provide users with answers that are strictly limited to the content of the uploaded drug label, thereby minimizing hallucinations and enhancing traceability. A user-friendly interface built with Streamlit allows users to upload FDA labeling PDFs and pose natural language queries. The chatbot extracts relevant sections using PyMuPDF and regex-based segmentation and generates responses constrained to those sections. To evaluate performance, semantic similarity scores were computed between generated answers and ground truth text using Sentence Transformers. Results across 10 breast cancer drug labels demonstrate high semantic alignment, with most scores ranging from 0.7 to 0.9, indicating reliable summarization and contextual fidelity. The chatbot achieved high semantic similarity scores (≥0.95 for concise sections) and ROUGE scores, confirming strong semantic and textual alignment. Comparative analysis with GPT-5-chat and NotebookLM demonstrated that our approach maintains accuracy and section-specific fidelity across models. The current work is limited to a small dataset, focused on breast cancer drugs. Future work will expand to diverse therapeutic areas and incorporate BERTScore and expert-based validation. Full article

AI-based educational chatbots can expand access to learning, but many remain limited to text-only interfaces and fixed infrastructures, while purely generative responses raise concerns of reliability and consistency. In this context, we present ROboMC, a portable and multimodal system that combines a validated knowledge base with generative responses (OpenAI) and voice–text interaction, designed to enable both text and voice interaction, ensuring reliability and flexibility in diverse educational scenarios. The system, developed in Django, integrates two response pipelines: local search using normalized keywords and fuzzy matching in the LocalQuestion database, and fallback to the generative model GPT-3.5-Turbo (OpenAI, San Francisco, CA, USA) with a prompt adapted exclusively for Romanian and an explicit disclaimer. All interactions are logged in AutomaticQuestion for later analysis, supported by a semantic encoder (SentenceTransformer—paraphrase-multilingual-MiniLM-L12-v2’, Hugging Face Inc., New York, NY, USA) that ensures search tolerance to variations in phrasing. Voice interaction is managed through gTTS (Google LLC, Mountain View, CA, USA) with integrated audio playback, while portability is achieved through deployment on a Raspberry Pi 4B (Raspberry Pi Foundation, Cambridge, UK) with microphone, speaker, and battery power. Voice input is enabled through a cloud-based speech-to-text component (Google Web Speech API accessed via the Python SpeechRecognition library, (Anthony Zhang, open-source project, USA) using the Google Web Speech API (Google LLC, Mountain View, CA, USA; language = “ro-RO”)), allowing users to interact by speaking. Preliminary tests showed average latencies of 120–180 ms for validated responses on laptop and 250–350 ms on Raspberry Pi, respectively, 2.5–3.5 s on laptop and 4–6 s on Raspberry Pi for generative responses, timings considered acceptable for real educational scenarios. A small-scale usability study (N ≈ 35) indicated good acceptability (SUS ~80/100), with participants valuing the balance between validated and generative responses, the voice integration, and the hardware portability. Although system validation was carried out in the eHealth context, its architecture allows extension to any educational field: depending on the content introduced into the validated database, ROboMC can be adapted to medicine, engineering, social sciences, or other disciplines, relying on ChatGPT only when no clear match is found in the local base, making it a scalable and interdisciplinary solution. Full article

With the rising incidence of traffic accidents and growing environmental concerns, the demand for advanced systems to ensure traffic and environmental safety has become increasingly urgent. This paper introduces an automated highway safety management framework that integrates computer vision and natural language processing for real-time monitoring, analysis, and reporting of traffic incidents. The system not only identifies accidents but also aids in coordinating emergency responses, such as dispatching ambulances, fire services, and police, while simultaneously managing traffic flow. The approach begins with the creation of a diverse highway accident dataset, combining public datasets with drone and CCTV footage. YOLOv11s is retrained on this dataset to enable real-time detection of critical traffic elements and anomalies, such as collisions and fires. A vision–language model (VLM), Moondream2, is employed to generate detailed scene descriptions, which are further refined by a large language model (LLM), GPT 4-Turbo, to produce concise incident reports and actionable suggestions. These reports are automatically sent to relevant authorities, ensuring prompt and effective response. The system’s effectiveness is validated through the analysis of diverse accident videos and zero-shot simulation testing within the Webots environment. The results highlight the potential of combining drone and CCTV imagery with AI-driven methodologies to improve traffic management and enhance public safety. Future work will include refining detection models, expanding dataset diversity, and deploying the framework in real-world scenarios using live drone and CCTV feeds. This study lays the groundwork for scalable and reliable solutions to address critical traffic safety challenges. Full article

The top shell, Turbo sazae, occurs commonly in the shallow rocky subtidal area of Jeju Island off the south coast of Korea, and it is one of the most valuable gastropod resources supporting the local shellfish industry. T. sazae landings in Jeju have declined dramatically in recent years, although the factors involved in this decline are yet to be identified. Recent studies also have reported that T. sazae is expanding its distribution range to the east coast of Korea, possibly due to the increasing seawater temperature. In this study, we investigated the hemocyte responses of T. sazae to elevated seawater temperatures in order to gain a better understanding of its immunological response to higher water temperatures. In this experiment, we exposed top shells to a gradual increase in seawater temperature, ranging from 22 °C to 30 °C, over a span of 9 days. We employed flow cytometry to assess various cellular immune responses, including hemocyte viability, phagocytosis capacity, and the production of reactive oxygen species (ROS) in T. sazae. The results showed that top shells exposed to elevated seawater temperature exhibited a significant decrease in phagocytosis capacity and an increase in ROS production after 3 days of the experiment. These findings indicate that an elevated seawater temperature imposes a stressful condition on T. sazae, characterized by reduced phagocytosis capacity and increased oxidative stress. Full article

Cumulative evidence from several pre-clinical studies suggests that restoration of plasma DNase activity in a thrombo-inflammatory state may improve clinical outcomes. Following injury, hyperactivated immune cells release large amounts of granular proteins together with DNA, which often accumulate in the surrounding environment in so-called neutrophil extracellular traps (NETs). Degradation of excess NETs by systemic DNase administration offers a promising therapeutic approach to ameliorate inflammation and dissolve intravascular clots. In order to expand the therapeutic utility of human DNase I, a variant of the enzyme was developed that has both a prolonged systemic half-life and a higher catalytic activity compared to Dornase alfa (Pulmozyme^®), the recombinant form of DNase I approved for inhaled therapy of cystic fibrosis. The hyperactive enzyme was “PASylated” by genetic fusion with a strongly hydrophilic and biodegradable PAS-polypeptide to increase its hydrodynamic volume and retard kidney filtration. A stable TurboCell™ CHO-K1-based cell line was generated which is suitable for the future production of PASylated DNase I according to good manufacturing practice (GMP). Furthermore, a robust bioprocess strategy was devised and an effective downstream process was developed. The final protein product is characterized by excellent purity, favorable physicochemical properties, a 14-fold higher DNA-degrading activity than Dornase alfa and a sustained pharmacokinetic profile, with a 22-fold slower clearance in rats. Full article

In this study, we conduct a pioneering and comprehensive examination of ChatGPT’s (GPT-3.5 Turbo) capabilities within the realm of Korean Grammatical Error Correction (K-GEC). Given the Korean language’s agglutinative nature and its rich linguistic intricacies, the task of accurately correcting errors while preserving Korean-specific sentiments is notably challenging. Utilizing a systematic categorization of Korean grammatical errors, we delve into a meticulous, case-specific analysis to identify the strengths and limitations of a ChatGPT-based correction system. We also critically assess influential parameters like temperature and specific error criteria, illuminating potential strategies to enhance ChatGPT’s efficacy in K-GEC tasks. Our findings offer valuable contributions to the expanding domain of NLP research centered on the Korean language. Full article

The scientific novelty of the proposed article lies in the development of the theory of analysis of the operation modes of group pump units operating on a long pipeline network with back pressure. This was achieved by creating electric equivalent circuits of group pump units based on the method of electrohydraulic analogy. Such equivalent circuits take into account pumps’ connection circuits, the configuration of the pipeline network, and the method of regulating technological parameters. A method for determining the characteristics of pump units with series and parallel connection is proposed. The dependences of the power consumed by the pump units on the change in the frequency of rotation of the adjustable pump are obtained for various parameters of the hydraulic network. This makes it possible to determine the limits of energy-efficient regulation of pump discharge. Analytical expressions for determining the lower limit of the pump rotation frequency for various circuits of turbo mechanism connection and various numbers of pumps operating simultaneously on a pipeline network with back pressure are proposed. The necessary range of adjustment of the pump rotation frequency with different circuits for turbo mechanism connection and different numbers of hydraulic machines operating at the same time is determined. The analysis of the obtained modes of pump units is performed and the possibility of expanding the controlled properties of the group electric drive systems of turbomachines when changing the direction of the rotation frequency and reversing the liquid in the event of an emergency situation is shown. Full article

Among recent state-of-the-art realistic image super-resolution (SR) intelligent algorithms, generative adversarial networks (GANs) have achieved impressive visual performance. However, there has been the problem of unsatisfactory perception of super-scored pictures with unpleasant artifacts. To address this issue and further improve visual quality, we proposed a perception-design-oriented PSRGAN with double perception turbos for real-world SR. The first-perception turbo in the generator network has a three-level perception structure with different convolution kernel sizes, which can extract multi-scale features from four $\frac{1}{4}$ size sub-images sliced by original LR image. The slice operation expands adversarial samples to four and could alleviate artifacts during GAN training. The extracted features will be eventually concatenated in later 3 × 2 upsampling processes through pixel shuffle to restore SR image with diversified delicate textures. The second-perception turbo in discriminators has cascaded perception turbo blocks (PTBs), which could further perceive multi-scale features at various spatial relationships and promote the generator to restore subtle textures driven by GAN. Compared with recent SR methods (BSRGAN, real-ESRGAN, PDM_SR, SwinIR, LDL, etc.), we conducted an extensive test with a $\times 4$ upscaling factor on various datasets (OST300, 2020track1, RealSR-Canon, RealSR-Nikon, etc.). We conducted a series of experiments that show that our proposed PSRGAN based on generative adversarial networks outperforms current state-of-the-art intelligent algorithms on several evaluation metrics, including NIQE, NRQM and PI. In terms of visualization, PSRGAN generates finer and more natural textures while suppressing unpleasant artifacts and achieves significant improvements in perceptual quality. Full article

The hydrogen hydrostatic thrust bearing (HHTB) is a key component of hydrogen liquefaction that impacts turbo-expander characteristics. To analyze the feasibility of using the HHTB in this application, characteristics of HHTBs were calculated using a CFD model. To upgrade the performance of the HHTB, the impacts of bearing structure and operating parameters on static performance were investigated. Dynamic characteristics of the HHTB were studied using the dynamic grid method. It was found that the load capacity of the HHTB is less than that of helium-lubricated bearings but higher than that of air- and methane-lubricated bearings. The turbulent kinetic energy of hydrogen is higher than that of other gases. Load capacity can be enhanced through boosting supplied pressure, expanding the diameter of supply orifices, reducing gas film clearance, increasing the orifices quantity and setting a circumferential groove. A reduction in disturbance amplitude slightly increased the bearing’s dynamic stiffness. The dynamic stability of the HHTB was improved by a small film clearance in response to disturbance. Full article

The article presents an overview of modern analytical methods and experimental studies on the use of heat exchangers as part of different schemes, as well as technologies that increase the efficiency of heat exchangers using renewable energy sources. The main types of heat exchangers, and the principles of their operation, are considered. In addition, modern technologies for increasing the efficiency of heat exchangers through design are described. The practical experience of using plate heat exchangers in industry has been studied. An overview of the software development that is used in the design and optimization of heat exchange devices, as well as for the improvement of their energy efficiency, is presented. The presented mathematical models can be used for software that is applicable both to individual segments of plates of heat exchangers and heat exchangers in general, taking into account the dependence of the installation of the entire circuit on environmental parameters and location. In conclusion, recommendations are given for further research directions in the field of using heat exchangers with the inclusion of renewable energy sources. The technique of an energy technology complex, including a heat pump, a photovoltaic panel, and a desalination plant, is presented. The methodology is built around the basic design and energy balance of the complex, and it is also considered from the point of view of the exergetic balance. This allows for the use of additional components, such as a turbo expander for the implementation of the organic Rankine cycle, a wind turbine, and a solar concentrator. This scientific approach can become unified for the design and operation of an energy technology complex. In addition, an exergetic calculation method is presented for a thermal desalination plant operating as part of an energy technology complex with renewable energy sources. Full article

This study designed an externally pressurized bearing and analyzed the rotordynamics of a turbo expander for a hydrogen liquefaction plant. The turbo expander, comprising a turbine and compressor wheel assembled to a shaft, lowered the temperature of the helium refrigerant. Its rated speed was 75,000 rpm, and an externally pressurized gas bearing was selected to support the rotor. Pressurized helium was used as the lubricant for the bearing operation. To design the rotor–bearing system, we conducted a bearing performance analysis and rotordynamic characteristic prediction using the developed numerical model. We calculated the bearing stiffness and flow rate of the bearing gas for various feed parameters and selected the appropriate orifice diameter for maximum stiffness. The predicted Campbell diagram showed that the system had a sufficient separation margin with the critical speed, and the predicted critical speed correlated well with the nonlinear orbit simulation. A successful operation was achieved with the manufactured turbo expander within the rated speed. The shaft vibration was monitored during the operation test, and the test results revealed two critical speeds below the rated speed, as predicted by the analytical model. In addition, the shaft vibration was maintained at <3 μm. Full article

This paper presents a numerical simulation analysis of the flow and combustion characteristics of a fuel-rich LOX (liquid oxygen)/kerosene gas generator in an ATR (air turbo rocket) engine, examining the effects of local parameters on the combustion flow field and performance. The analysis considers variations in unit injector mixing ratios and unit mass flow rates. The results indicate that as the mixing ratio in the inner-ring injectors increases (while the mixing ratio in the middle-ring injectors decreases), the oxygen concentration area near the axis zone and the 50% radius zone of the gas generator expands. Conversely, the kerosene concentration area near the axis zone decreases while gradually increasing near the 50% radius zone. In the flow direction section, there is an inverse relationship between the variation trend of local temperature and the oxygen concentration in the local area. As the oxygen concentration increases, the temperature decreases. The temperature distribution across the cross-section of the gas generator exhibits a circular pattern. When the mixing ratio (or mass flow rates) of the unit injector are perfectly balanced, the temperature distribution becomes highly uniform. A larger disparity in flow rate between the inner ring injector and the middle ring injector leads to a lower combustion efficiency. This effect differs from the effect of the mixing ratio difference between the two injector rings. Increasing the mixing ratio in the inner-ring injectors (or decreasing the mixing ratio in the middle-ring injectors) initially leads to a rise in combustion efficiency, followed by a subsequent decline. The maximum combustion efficiency of 89.10% is achieved when the mixing ratio is set to Km_-1 = 0.7 and Km_-2 = 2.79, respectively. Increasing the flow rate in the inner-ring injectors (or decreasing the flow rate in the middle-ring injectors) initially leads to an improvement in combustion efficiency, followed by a subsequent reduction. The maximum combustion efficiency of 86.13% is achieved when the mass flow rate is set to m_-1 = m_-2 = 0.1 kg/s. Full article

This paper provides a simple methodology for the design of the air reactor of a chemical looping combustor to optimize its characteristics when it is employed connected to a turbo expander to produce power. The design process, given a certain objective (e.g., electric power) defines the reactor specifics, namely height and diameter, taking into account the following aspects: solids inventory of the air reactor; gas velocity; air reactor transport disengaging height (TDH); solids concentration profile along the reactor height, dense bed height; freeboard height; pressure drop depending on air reactor injectors design and configuration. The total air reactor height was about 9.5 m, while the diameter was about 1.8 m. The total inventory was about 10,880 kg; while the circulation rate in the air reactor was about 110 kg/s. The operating pressure and temperature were, respectively, 12 bar and 1200 °C. The average velocity of the gases inside the reactor was about 4 m/s. The fluidization regime resulted to be comprised between turbulent and fast fluidization. Further work must be directed into the estimate of the pressure drop of the reactor, which will affect the plant efficiency in a considerable way. Full article

Worldwide concerns over energy optimization and harnessing have sparked a variety of research on utilizing wasted energy. Natural gas, in particular, has been an essential energy commodity for domestic and industrial applications. This paper approaches the same principle throughout the Egyptian natural gas conventional national grid. The real case of Egypt’s Al-Alamein city gas line extension is analytically studied for energy harness. Thus, this paper aims to study the replacement of conventional energy-wasting pressure-reducing stations with power-producing expander generator systems to utilize pressure reduction in electricity generation. The mathematical equations that describe the energy recovery processes are presented. Reliable commercial software is used to verify the possible implementation of turbo expander stations in three scenarios. Each scenario is defined, and its main components and numerical parameters are introduced. Hence, the resulting energy outputs and applicability of all scenarios are compared to provide adequate energy recovery solutions. Full article