Association rule mining is a fundamental data mining technique for uncovering latent relationships among variables in large-scale datasets. However, conventional approaches rely on single-metric filtering strategies, which are insufficient for capturing the inherent multi-criteria nature of rule quality. To address this limitation, this study formulates ARM as a many-objective optimization problem and proposes a hybrid algorithm, NSGA-II/DE-ARM, that simultaneously optimizes four rule-quality measures: support, confidence, lift, and NetConf. The proposed algorithm enhances the NSGA-II framework by integrating binary differential evolution operators, an adaptive operator selection mechanism, lift-weighted tournament selection, and a constraint-domination principle combined with a dynamic minimum support threshold. Its performance was evaluated using two datasets: a SIPRI-World Bank panel dataset consisting of defense industry and macroeconomic indicators covering 46 items over the 2002-2023 period, and the UCI Mushroom benchmark dataset consisting of 118 items. Across 30 independent runs on the SIPRI-World Bank dataset, NSGA-II/DE-ARM outperformed the Apriori baseline in all four metrics (mean lift = 4.748, confidence = 0.853, support = 0.146, NetConf = 0.789), with large effect sizes (Cohen's d = 1.77-5.77, p < 0.001 in each case). On the Mushroom benchmark dataset, the proposed method also achieved substantial improvements, with Cohen's d values ranging from 0.93 to 6.16. NSGA-II/DE-ARM generated 68 Pareto-optimal rules in a representative run and achieved the highest hypervolume values on both datasets, with HV = 3.231 for SIPRI-World Bank and HV = 6.262 for Mushroom. These results suggest that NSGA-II/DE-ARM offers decision-makers a broader and more balanced multi-criteria solution set than single-metric filtering approaches.
Optimizing flapping-foil kinematics for underwater propulsion is challenging due to strong temporal coupling and nonlinear fluid-structure interactions. Most existing approaches rely on parameterized motion profiles, which restrict the accessible design space. A non-parametric kinematic optimization framework based on the discrete adjoint method is developed, enabling direct optimization of time-resolved motions without predefined functional forms. A Morison-based low-order hydrodynamic model, calibrated against Computational Fluid Dynamics (CFD), is employed for efficient evaluation within a validated regime. Results show that optimized motions substantially enhance propulsion performance over conventional sinusoidal motions, yielding non-sinusoidal, high-efficiency kinematics. In thrust-maximization cases, the optimized kinematics achieve a 50.29% increase in mean thrust by redistributing heave and pitch amplitudes and timing. Under balanced thrust-power conditions, the optimized motions consistently outperform sinusoidal counterparts. In power-minimization cases, a "generator-like" regime emerges, indicating a reversal of net energy transfer enabled by the non-parametric formulation. These results demonstrate that non-parametric optimization provides enhanced design flexibility and improved propulsion performance, offering a practical framework for biomimetic underwater propulsion design.
Trust in embodied intelligence is dynamic, contextual, and interaction-dependent, but many existing computational approaches still model trust using static similarity structures. This study proposes and evaluates a compositional trust modeling approach based on quantum natural language processing (QNLP). Using open-ended survey responses about human trust in embodied agents, we compared classical NLP clustering and QNLP-based clustering in terms of dimension coverage, semantic coherence, contextual sensitivity, and robustness. The QNLP pipeline captured richer latent structure, producing ten clusters and identifying eight trust dimensions, including two emergent dimensions: calibrated trust and predictive reliability. Compared with classical approaches, QNLP clusters showed improved semantic separation and stronger context retention under preprocessing variation. These findings support a temporally structured view of trust in embodied AI and demonstrate that compositional quantum-inspired representations can reveal nuanced trust dynamics that are difficult to detect with conventional methods. This study contributes both a methodological framework for trust-sensitive text modeling and a theoretical account linking trust formation to retrospective calibration and prospective expectation in human-agent interaction.
Resolving the kinematic redundancy of 7-DoF humanoid arms to generate natural, human-like motions remains a fundamental challenge in biomimetic robotics. This paper presents a hybrid inverse kinematics (IK) framework that learns a pose-dependent redundancy parameter and integrates it into a differential IK solver. Specifically, we employ the stereographic Shoulder-Elbow-Wrist (SEW) angle as a well-conditioned geometric parameterization. This formulation transforms the algorithmic singularity into a unidirectional half-line, which can be oriented outside the typical reachable workspace. To specify the optimal configuration within the self-motion manifold, a motion dataset was collected by teleoperating a humanoid arm via an anthropomorphic wearable exoskeleton. This approach translates operator-specific postural preferences into the robot's joint space. A lightweight neural network was then trained to learn the mapping from end-effector poses to these operator-specific SEW angles. By incorporating the predicted SEW angle as a dynamic secondary objective in the null space of the primary tracking task, the proposed framework enables natural redundancy resolution while preserving end-effector tracking accuracy. Both simulations and real-robot experiments were conducted to validate the approach. Results show that, compared to the average performance of static fixed-parameter strategies, the proposed method improves the Joint Configuration Quality Index (CQI) by 22.5% and reduces energy costs by 11.3%. Moreover, the sub-millisecond inference latency (0.44 ms) facilitates seamless integration into real-time control pipelines.
Effective diabetes management heavily relies on appropriate insulin administration, which strongly depends on the correct administration strategy. In this sense, insulin administration plays a fundamental role, as its use depends on the patient's clinical condition and diabetes type. Traditional syringe-based methods require proper training to ensure that insulin is successfully delivered into the subcutaneous tissue, where it can be absorbed and metabolized; however, it is desirable to develop an insulin applicator that does not require training for its appropriate use. Aiming to provide support solutions that help patients to develop a correct administration technique, a biodesign-based methodology, coupled with biomimetic concepts, is employed to design a device that assists the user in creating a stable skin fold and guiding needle orientation during injection without requiring exhaustive training for device usage. A three-step approach is employed for the design, where computational fluid dynamics (CFD) and finite element analysis (FEA) methods are employed to ensure that the device produces a laminar insulin flow and the device strength is tested. It should be pointed out both methods are required since complications produced by sudden flows must be avoided, with CFD allowing assessment of the device mechanical properties in terms of the device strength. Initial functional evaluation indicates that the proposed approach does not require extensive training or complex operational procedures, facilitating its integration into everyday use. The device design is validated from the results obtained for the CFD analysis, as no turbulent flow is produced, whereas the FEA indicates that the geometrical form can handle the stresses produced by the folding generation without generating excessive deformations. Moreover, an infrared thermography analysis is also carried out to find out if the folding force generation is located in the zone of interest, the results of which indicate that the device operates in the desired physical zone.
