Publications

Gulnoza Adilmetova, Ruslan Nassyrov, Aizhan Meyerbekova, Aknur Karabay, Huseyin Atakan Varol, Mei-Yen Chan, Evaluating ChatGPT’s Multilingual Performance in Clinical Nutrition Advice Using Synthetic Medical Text: Insights from Central Asia, The Journal of Nutrition, Volume 155, Issue 3, 2025, Pages 729-735, ISSN 0022-3166, https://doi.org/10.1016/j.tjnut.2024.12.018. (https://www.sciencedirect.com/science/article/pii/S0022316624012458) Abstract: Background Although large language models like ChatGPT-4 have demonstrated competency in English, their performance for minority groups speaking underrepresented languages, as well as their ability to adapt to specific sociocultural nuances and regional cuisines, such as those in Central Asia (for example, Kazakhstan), still requires further investigation. Objectives To evaluate and compare the effectiveness of the ChatGPT-4 system in providing personalized, evidence-based nutritional recommendations in English, Kazakh, and Russian in Central Asia. Methods This study was conducted from 15 May to 31 August, 2023. On the basis of 50 mock patient profiles, ChatGPT-4 generated dietary advice, and responses were evaluated for personalization, consistency, and practicality using a 5-point Likert scale. To identify significant differences between the 3 languages, the Kruskal–Wallis test was conducted. Additional pairwise comparisons for each language were carried out using the post hoc Dunn's test. Results ChatGPT-4 showed a moderate level of performance in each category for English and Russian languages, whereas in Kazakh language, outputs were unsuitable for evaluation. The scores for English, Russian, and Kazakh were as follows: for personalization, 3.32 ± 0.46, 3.18 ± 0.38, and 1.01 ± 0.06; for consistency, 3.48 ± 0.43, 3.38 ± 0.39, and 1.09 ± 0.18; and for practicality, 3.25 ± 0.41, 3.37 ± 0.38, and 1.07 ± 0.15, respectively. The Kruskal–Wallis test indicated statistically significant differences in ChatGPT-4's performance across the 3 languages (P < 0.001). Subsequent post hoc analysis using Dunn’s test showed that the performance in both English and Russian was significantly different from that in Kazakh. Conclusions Our findings show that, despite using identical prompts across 3 distinct languages, the ChatGPT-4's capability to produce sensible outputs is limited by the lack of training data in non-English languages. Thus, a customized large language model should be developed to perform better in underrepresented languages and to take into account specific local diets and practices. Keywords: AI; LLMs; personalized diet; precision nutrition; Chatbots; Medical Natural Language Processing (NLP)

@ARTICLE{10843698, author={Balapan, Azat and Yeralkhan, Rauan and Aryslanov, Alikhan and Kalimuldina, Gulnur and Yeshmukhametov, Azamat}, journal={IEEE Access}, title={A Novel Pattern Recognition Method for Self-Powered TENG Sensor Embedded to the Robotic Hand}, year={2025}, volume={13}, number={}, pages={14101-14112}, keywords={Robot sensing systems;Hands;Thumb;Shape;Sensors;Data collection;Tactile sensors;Data visualization;Convolutional neural networks;Carbon;Dataset collection;machine learning;robot hand design;signal processing;TENG sensor}, doi={10.1109/ACCESS.2025.3530465}}

@ARTICLE{10697322, author={Moger, Gourav and Varol, Huseyin Atakan}, journal={IEEE/ASME Transactions on Mechatronics}, title={Design and Implementation of a Mobile Robot With Variable-Diameter Wheels}, year={2024}, volume={}, number={}, pages={1-10}, keywords={Wheels;Robots;Gears;Legged locomotion;Robot kinematics;Kinematics;Motors;Thermal stability;Deformation;Complexity theory;Hybrid wheeled mobile robots (WMRs);kinematic modeling;variable-diameter wheels}, doi={10.1109/TMECH.2024.3457011}}

@Article{s24185993, AUTHOR = {Rakhimzhanova, Tomiris and Kuzdeuov, Askat and Varol, Huseyin Atakan}, TITLE = {AnyFace++: Deep Multi-Task, Multi-Domain Learning for Efficient Face AI}, JOURNAL = {Sensors}, VOLUME = {24}, YEAR = {2024}, NUMBER = {18}, ARTICLE-NUMBER = {5993}, URL = {https://www.mdpi.com/1424-8220/24/18/5993}, PubMedID = {39338738}, ISSN = {1424-8220}, ABSTRACT = {Accurate face detection and subsequent localization of facial landmarks are mandatory steps in many computer vision applications, such as emotion recognition, age estimation, and gender identification. Thanks to advancements in deep learning, numerous facial applications have been developed for human faces. However, most have to employ multiple models to accomplish several tasks simultaneously. As a result, they require more memory usage and increased inference time. Also, less attention is paid to other domains, such as animals and cartoon characters. To address these challenges, we propose an input-agnostic face model, AnyFace++, to perform multiple face-related tasks concurrently. The tasks are face detection and prediction of facial landmarks for human, animal, and cartoon faces, including age estimation, gender classification, and emotion recognition for human faces. We trained the model using deep multi-task, multi-domain learning with a heterogeneous cost function. The experimental results demonstrate that AnyFace++ generates outcomes comparable to cutting-edge models designed for specific domains.}, DOI = {10.3390/s24185993} }

@ARTICLE{10509679, author={Keutayeva, Aigerim and Abibullaev, Berdakh}, journal={IEEE Access}, title={Data Constraints and Performance Optimization for Transformer-Based Models in EEG-Based Brain-Computer Interfaces: A Survey}, year={2024}, volume={12}, number={}, pages={62628-62647}, keywords={Brain modeling;Electroencephalography;Transformers;Data models;Decoding;Task analysis;Training;Brain-computer interfaces;data scarcity;survey;deep learning;transformer;EEG;self-attention mechanism;motor imagery;BCI}, doi={10.1109/ACCESS.2024.3394696}}

@inproceedings{maxutov-etal-2024-llms, title = "Do {LLM}s Speak {K}azakh? A Pilot Evaluation of Seven Models", author = "Maxutov, Akylbek and Myrzakhmet, Ayan and Braslavski, Pavel", editor = {Ataman, Duygu and Derin, Mehmet Oguz and Ivanova, Sardana and K{\"o}ksal, Abdullatif and S{\"a}lev{\"a}, Jonne and Zeyrek, Deniz}, booktitle = "Proceedings of the First Workshop on Natural Language Processing for Turkic Languages (SIGTURK 2024)", month = aug, year = "2024", address = "Bangkok, Thailand and Online", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.sigturk-1.8", pages = "81--91", abstract = "We conducted a systematic evaluation of seven large language models (LLMs) on tasks in Kazakh, a Turkic language spoken by approximately 13 million native speakers in Kazakhstan and abroad. We used six datasets corresponding to different tasks {--} questions answering, causal reasoning, middle school math problems, machine translation, and spelling correction. Three of the datasets were prepared for this study. As expected, the quality of the LLMs on the Kazakh tasks is lower than on the parallel English tasks. GPT-4 shows the best results, followed by Gemini and . In general, LLMs perform better on classification tasks and struggle with generative tasks. Our results provide valuable insights into the applicability of currently available LLMs for Kazakh. We made the data collected for this study publicly available: https://github.com/akylbekmaxutov/LLM-eval-using-Kazakh.", }

@Article{buildings14040935, AUTHOR = {Karaca, Ferhat and Tleuken, Aidana and Pineda-Martos, Rocío and Ros Cardoso, Sara and Orel, Daniil and Askar, Rand and Agibayeva, Akmaral and Güemez, Elena Goicolea and Salles, Adriana and Varol, Huseyin Atakan and Braganca, Luis}, TITLE = {Cultivating Sustainable Construction: Stakeholder Insights Driving Circular Economy Innovation for Inclusive Resource Equity}, JOURNAL = {Buildings}, VOLUME = {14}, YEAR = {2024}, NUMBER = {4}, ARTICLE-NUMBER = {935}, URL = {https://www.mdpi.com/2075-5309/14/4/935}, ISSN = {2075-5309}, ABSTRACT = {Due to its intricate production processes, complex supply chains, and industry-specific characteristics, the construction industry faces unique challenges in adopting circular economy (CE) principles that promote resource equity. To address this issue, this study aims to delve into identifying stakeholders’ opinions and perceptions regarding key CE strategies across different stages of the building life cycle (BLC). Both European and non-European stakeholders within the “CircularB” COST Action network and beyond participated in this research. Three methods were employed to assess stakeholders’ opinions: an online survey, a structured survey with a semi-guided workshop, and creative thinking round table discussions. Natural language processing (NLP), specifically topic modelling and sentiment analysis, was used to analyse the data collected from the online survey, which gathered text-based opinions from 209 participants on the cost-benefit aspects of circularity strategies. The structured survey, which collected data from 43 workshop participants, evaluated the perceived importance of CE strategies across various BLC phases and assessed the adoption of selected CE strategies in current or past projects. Finally, the Six Thinking Hats® activity, employed in the round table discussions, generated ideas from 25 professionals regarding the broader implementation challenges and opportunities of CE in construction. The research findings highlight the need to bridge the gap between theory and practice by fostering active industry stakeholder involvement in the transition to a CE model. The analyses of the collected stakeholder opinions through the three activities contribute to proactive and collaborative efforts aimed at advancing resource equity in the construction sector and promoting just and inclusive resource use. In summary, this research offers a comprehensive understanding of stakeholders’ opinions on CE strategies and provides guidance for the development of targeted policies and strategies to accelerate the integration of CE principles in the construction industry.}, DOI = {10.3390/buildings14040935} }

@article{https://doi.org/10.1002/eem2.12808, author = {Mubarak, Assem and Sarsembayev, Bayandy and Serik, Yerzhigit and Onabek, Abdirakhman and Kappassov, Zhanat and Bakenov, Zhumabay and Tsuchiya, Kazuyoshi and Kalimuldina, Gulnur}, title = {Quenched PVDF/PMMA Porous Matrix for Triboelectric Energy Harvesting and Sensing}, journal = {ENERGY \& ENVIRONMENTAL MATERIALS}, volume = {n/a}, number = {n/a}, pages = {e12808}, keywords = {N,N′-poly(methyl methacrylate) (PMMA), robotic tactile sensors, polyvinylidene fluoride (PVDF), quenching, triboelectric nanogenerators (TENGs)}, doi = {https://doi.org/10.1002/eem2.12808}, url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/eem2.12808}, eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1002/eem2.12808}, note = {e12808 EEM-2024-0184.R1}, abstract = {The rapid development of nanotechnology has significantly revolutionized wearable electronics and expanded their functionality. Through introducing innovative solutions for energy harvesting and autonomous sensing, this research presents a cost-effective strategy to enhance the performance of triboelectric nanogenerators (TENGs). The TENG was fabricated from polyvinylidene fluoride (PVDF) and N, N′-poly(methyl methacrylate) (PMMA) blend with a porous structure via a novel optimized quenching method. The developed approach results in a high β-phase content (85.7\%) PVDF/3wt.\%PMMA porous blend, known for its superior piezoelectric properties. PVDF/3wt.\%PMMA modified porous TENG demonstrates remarkable electrical output, with a dielectric constant of 40 and an open-circuit voltage of approximately 600 V. The porous matrix notably increases durability, enduring over 36 000 operational cycles without performance degradation. Moreover, practical applications were explored in this research, including powering LEDs and pacemakers with a maximum power output of 750 mW m−2. Also, TENG served as a self-powered tactile sensor for robotic applications in various temperature conditions. The work highlights the potential of the PVDF/PMMA porous blend to utilize the next-generation self-powered sensors and power small electronic devices.}}

@article{https://doi.org/10.1111/jiec.13502, author = {Tleuken, Aidana and Orel, Daniil and Iskakova, Anel and Varol, Huseyin Atakan and Karaca, Ferhat}, title = {Exploring stakeholders’ opinions on circular economy in the construction sector: A natural language processing analysis of social media discourse}, journal = {Journal of Industrial Ecology}, volume = {n/a}, number = {n/a}, pages = {}, keywords = {circular economy, construction sector, industrial ecology, sentiment analysis, sustainable buildings, topic modeling}, doi = {https://doi.org/10.1111/jiec.13502}, url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/jiec.13502}, eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1111/jiec.13502}, abstract = {Abstract The construction industry has been criticized for its negative environmental impacts, leading industry experts to advocate for a shift toward a circular economy (CE) model. However, there is a lack of research on stakeholder opinions regarding that. This research paper examines stakeholders’ perspectives on implementing CE principles in the construction industry by conducting artificial intelligence-powered natural language processing (NLP) through online sources. It answers three questions: What themes and concepts are associated with the CE in construction? How do opinions on the CE vary across different online platforms? And what factors shape positive attitudes toward the CE? The data obtained from various platforms showed that 57\% of sentiments were positive, 28\% were neutral, and 15\% were negative. This research provides critical knowledge on the analysis of CE representation on social media in construction. Moreover, a webpage tool has been created that can assess any input opinion on the scale (positive, neutral, or negative) for further use (https://ce-sentiment.streamlit.app/). This NLP-based research of social media discourse in the construction sector can directly influence policy decisions by offering real-time insights into public sentiment and preferences, shaping regulations that align with societal needs. It also provides industry professionals with data-driven guidance, enabling them to identify growth opportunities and innovation pathways within the CE, ultimately fostering a more sustainable and prosperous future.} }

@InProceedings{10.1007/978-3-031-57800-7_31, author="Karaca, Ferhat and Tleuken, Aidana and Pineda-Martos, Roc{\'i}o and Cardoso, Sara Ros and Askar, Rand and Salles, Adriana and G{\"u}emez, Elena Goicolea and Agibayeva, Akmaral and Varol, Huseyin Atakan and Braganca, Luis", editor="Ungureanu, Viorel and Bragan{\c{c}}a, Lu{\'i}s and Baniotopoulos, Charalambos and Abdalla, Khairedin M.", title="Analysing Stakeholder Opinions Within the COST Action CA21103 CircularB and Beyond: Circular Economy Implementation in Construction", booktitle="4th International Conference ``Coordinating Engineering for Sustainability and Resilience'' {\&} Midterm Conference of CircularB ``Implementation of Circular Economy in the Built Environment''", year="2024", publisher="Springer Nature Switzerland", address="Cham", pages="335--345", abstract="This study examines the importance and adoption of circular economy (CE) principles within the construction industry by focusing on stakeholders' opinions on key CE strategies across different building life cycle stages. The study draws insights from the perceptions of European-based stakeholders who actively participated in the CircularB Workshop 1 Part 2, entitled Creating a Roadmap towards Circularity in the Built Environment - State-of-the-Art. The research comprises two parts. In the first part, a structured survey was employed to systematically collect opinions on the levels of awareness and variations among the adoption and importance of selected CE implementation strategies within the construction sector. The second part engaged stakeholders in a dynamic creative thinking activity, posing seven targeted questions allowing participants to offer multiple answers for each query. Overall, the study sheds light on the multifaceted challenges and opportunities inherent in fostering CE within the construction domain by highlighting the significance of recognising and addressing systemic barriers within the CE framework, the importance of product design for disassembly, and the efficient production of reusable and recyclable materials. Furthermore, it emphasises the necessity to motivate industry stakeholders to participate actively in the transition to a CE, bridging the gap between theory and practice frameworks and increasing the engagement of policymakers and governments.", isbn="978-3-031-57800-7" }