KAN Research Institute

Artificial Intelligence (AI) is revolutionizing reproductive medicine by enhancing fertility treatments, childbirth monitoring, and postnatal care. AI-driven technologies, including machine learning (ML), natural language processing (NLP), and robotic-assisted surgery, improve diagnostic accuracy, optimize treatment plans, and support clinical decision-making in assisted reproductive technology (ART). While AI has demonstrated significant potential, challenges such as ethical concerns, data privacy risks, algorithmic biases, and regulatory uncertainties remain critical barriers to widespread adoption in reproductive healthcare. This chapter explores the role, advancements, and challenges of AI in reproductive medicine, highlighting its impact on fertility, pregnancy, and postnatal care. A comprehensive analysis of AI applications in reproductive healthcare, including embryo selection, sperm and oocyte quality assessment, ovulation tracking, prenatal diagnostics, labor monitoring, and postpartum depression detection. AI aids in early detection of obstetric complications, optimizes neonatal care in intensive care units (NICUs), and enhances fertility treatments through predictive analytics. AI-driven virtual assistants support maternal health, while machine learning-based tools improve diagnostic precision and clinical decision-making. Addressing ethical and regulatory concerns is essential for equitable AI integration in reproductive medicine. Future advancements require interdisciplinary collaboration, standardized validation through randomized controlled trials, and responsible AI deployment to enhance patient outcomes while maintaining clinical expertise and ethical integrity.

Self-propulsion is a quintessential aspect of biological systems, which can induce nonequilibrium phenomena that have no counterparts in passive systems. Motivated by biophysical interest together with recent advances in experimental techniques, active matter has been a rapidly developing field in physics. Meanwhile, over the past few decades, topology has played a crucial role to understand certain robust properties appearing in condensed matter systems. For instance, the nontrivial topology of band structures leads to the notion of topological insulators, where one can find robust gapless edge modes protected by the bulk band topology. We here review recent progress in an interdisciplinary area of research at the intersection of these two fields. Specifically, we give brief introductions to active matter and band topology in Hermitian systems, and then explain how the notion of band topology can be extended to nonequilibrium (and thus non-Hermitian) systems including active matter. We review recent studies that have demonstrated the intimate connections between active matter and topological materials, where exotic topological phenomena that are unfeasible in passive systems have been found. A possible extension of the band topology to nonlinear systems is also briefly discussed. Active matter can thus provide an ideal playground to explore topological phenomena in qualitatively new realms beyond conservative linear systems.

In 2017, a large outbreak of Legionnaires' disease, involving 58 patients, occurred in a public bath facility in Hiroshima Prefecture, Japan. We analyzed 94 Legionella pneumophila strains isolated from patients and the public bath facility using molecular typing methods, including Pulsed-Field Gel Electrophoresis, sequence-based typing, multi-locus variable number tandem repeat analysis, and Whole-Genome Sequencing (WGS). Genotypes obtained using these molecular epidemiological typing methods were highly correlated with each other. L. pneumophila strains of various genotypes were isolated from the public bath facility, of which only ST2398 and ST2399 were isolated from patients. ST2398 and ST2399, isolated from patients, bath water, and swabs, derived from one common circulating system at the bath facility out of seven were found to be novel genotypes and a highly clonal genetic lineage by single nucleotide variant (SNV) analysis based on WGS. The result of haplotype network analysis based on SNVs showed that ST2398 and ST2399 differed only approximately 30-42 SNVs, and some environmental strains that differed by only 0-3 SNVs from patient strains were isolated. These results demonstrated that this outbreak was caused by L. pneumophila assigned to the ST2398 and ST2399 clades. We found that at least three patients were co-infected with different clusters of L. pneumophila serogroup 1. Our results show that several strains must be isolated from a single sample to consider the accumulation of mutations in water and co-infection when investigating outbreaks.

In this study static load testing of three full scale drilled piles with combined tip and shaft grouting was conducted, where pile head load, pile head settlement, and compression strain of the piles were recorded. It was found that post-grouting tip and shaft of drilled piles can increase pile capacity significantly. Strain incompatibility existed between side friction and tip bearing of the post-grouted piles as the side friction was mobilized before the tip bearing was activated and the piles failed due to the plunge failure before the side friction reached the ultimate strength. It is found that the side shear enhancement of the same type soil varied based on the depth of the soil layer and decreased from pile head toward the tip of piles. The experimental program and theoretical analysis are presented in this paper.

To address the challenges of user site aggregation and noise accumulation interference in multi-user time synchronization, this study proposes an optical fiber time transfer system based on a star topology. The system distributes time and frequency signals to multiple slave users through a master terminal device combined with Time Division Multiple Access (TDMA) technology and optical splitters. Each slave site implements compensation within its time transfer equipment to synchronize its local time signal with the master reference signal. Through TDMA-based two-way time comparison, the system ensures the mutual independence of slave devices while eliminating multi-user signal noise accumulation. Experimental results show that all slave devices achieve pslevel stability and a time difference of less than 100 ps with no cross-interference. The root mean square (RMS) and standard deviation (STD) are less than 10 ps. This scheme provides a scalable solution and feasibility for distributed user nodes requiring high-precision time synchronization.

Early and accurate prediction of diabetes is one of the dire objectives of medical diagnostics. This paper contains a comparative work of the K-Nearest Neighbor (KNN) and Decision Tree Classifier (DTC) in predicting diabetes based on clinical health indicators. A critical preprocessing of the dataset was done taking care of missing values, feature scaling, and normalization to guarantee the integrity and soundness of the models. Each of these two algorithms was run 10 times to check the consistency and performance. As the performance measure showed, KNN seemed to give better results than DTC with an average accuracy of around 92.5 percent, whereas Decision Tree got 90.0 percent. The worst accuracy achieved on experimentation with KNN was 87.5 which shows consistency in prediction ability. Besides, an average score of 0.91 of KNN compared to 0.80 of DTC highlights the use of KNN as a more precise and robust algorithm to work with the standardized data. Standard error bars also indicate a less varying KNN results. The conclusion of this comparative study is that both models are suitable to use in classifying diabetes but KNN is more reliable and accurate and therefore is a good option to use in any situation that involves high performance diagnosis using health indicators as the input. The results of the study could help practitioners and developers to select the proper machine learning models to integrate with clinical decision support systems.

Migraine is among the most frequently observed neurovascular malady affecting almost 1billion people globally and is defined by frequent episodes of headache, accompanying neurological problems, and a markedly reduced standard of health. Migraine pathogenesis is directly linked to the CGRP (Calcitonin-gene-related-peptide), a strong vasodilating protein entangled in the exchange of nociceptor signals, and the cyclo-oxygenase (COX) enzymes, influencing inflammatory responses which aggravate migraine manifestations. Although COX inhibiting drugs as well as CGRP receptor blockers have therapeutical advantages when administered independently, however, they also have drawbacks, such as insufficient effectiveness or side effects, highlighting the necessity of synergistic approaches. In this regard, SYMBRAVO is a unique combinatorial pharmacological approach that aims to combine the advantages of COX-2 suppression and CGRP regulation. We have explored themolecular basis, pharmacological interaction, and clinical effectiveness of SYMBRAVO in this study, with a focus on how it can improve effectiveness, reduce adverse reactionsand overcome barriers to single drug therapy. The results support a paradigm change towards integrated migraine therapy strategies that emphasize controlled, multi-pathway regulation.

In the field of refuse collection, the need for contactless waste collection is increasing due to infectious diseases and labor shortage. In this study, we developed an autonomous robot that can collect and automatically transport waste containers. In this paper, we examine the traffic rules and travel area according to the road environment on which the robot runs and conduct a travel test using a fleet control system that controls multiple robots in a simulated environment. The results suggest that the success rate of autonomous driving of robots increases when the driving area and route are set according to the road environment and “waiting points” are set at locations where robots pass each other, and that robots can execute tasks without interfering with each other by sharing their location information and planning their routes. The results also suggest that robots can execute tasks without interfering with each other by sharing their location information and planning their paths.

This study aims to investigate the satisfaction levels of the elderly concerning their physical well-being and sociocultural entertainment needs, with a particular focus on engagement in sensor-based gaming. The objective is to systematically synthesise and evaluate empirical research findings regarding the factors influencing accessibility and enthusiasm for sensor-based gaming among older adults. Employing the PRISMA methodology, a comprehensive review was conducted, resulting in the identification of 43 English-language scholarly sources from the Web of Science, Scopus, and IEEE Xplore, covering the period from 2019 to 2024. The findings reveal that critical determinants affecting elderly engagement in sensor-based gaming include physical activity levels, cognitive capabilities, game incentive mechanisms, cooperative-competitive dynamics, personalised feature design, group participation, and gaming frequency. These insights contribute significantly to the discourse on utilizing exergames to enhance the physical and psychological well-being of the ageing population, providing valuable implications for researchers and policymakers aiming to address the challenges associated with an ageing demographic.

Endothelial Notch signaling is critical for tumor angiogenesis. Notch1 blockade can interfere with tumor vessel function but causes tissue hypoxia and gastrointestinal toxicity. Notch4 is primarily expressed in endothelial cells, where it may promote angiogenesis; however, effective therapeutic targeting of Notch4 has not been successful. We developed highly specific Notch4-blocking antibodies, 6-3-A6 and humanized E7011, allowing therapeutic targeting of Notch4 to be assessed in tumor models. Notch4 was expressed in tumor endothelial cells in multiple cancer models, and endothelial expression was associated with response to E7011/6-3-A6. Anti-Notch4 treatment significantly delayed tumor growth in mouse models of breast, skin, and lung cancers. Enhanced tumor inhibition occurred when anti-Notch4 treatment was used in combination with chemotherapeutics. Endothelial transcriptomic analysis of murine breast tumors treated with 6-3-A6 identified significant changes in pathways of vascular function but caused only modest change in canonical Notch signaling. Analysis of early and late treatment timepoints revealed significant differences in vessel area and perfusion in response to anti-Notch4 treatment. We conclude that targeting Notch4 improves tumor growth control through endothelial intrinsic mechanisms.Significance:A first-in-class anti-Notch4 agent, E7011, demonstrates strong antitumor effects in murine tumor models including breast carcinoma. Endothelial Notch4 blockade reduces perfusion and vessel area.