Hasil untuk "Machine design and drawing"

Mariëlle Creusen, J. Schoormans

A. Mesaros, Toni Heittola, Emmanouil Benetos et al.

Chengxu Yang, Qipeng Wang, Mengwei Xu et al.

Federated learning (FL) is an emerging, privacy-preserving machine learning paradigm, drawing tremendous attention in both academia and industry. A unique characteristic of FL is heterogeneity, which resides in the various hardware specifications and dynamic states across the participating devices. Theoretically, heterogeneity can exert a huge influence on the FL training process, e.g., causing a device unavailable for training or unable to upload its model updates. Unfortunately, these impacts have never been systematically studied and quantified in existing FL literature. In this paper, we carry out the first empirical study to characterize the impacts of heterogeneity in FL. We collect large-scale data from 136k smartphones that can faithfully reflect heterogeneity in real-world settings. We also build a heterogeneity-aware FL platform that complies with the standard FL protocol but with heterogeneity in consideration. Based on the data and the platform, we conduct extensive experiments to compare the performance of state-of-the-art FL algorithms under heterogeneity-aware and heterogeneity-unaware settings. Results show that heterogeneity causes non-trivial performance degradation in FL, including up to 9.2% accuracy drop, 2.32 × lengthened training time, and undermined fairness. Furthermore, we analyze potential impact factors and find that device failure and participant bias are two potential factors for performance degradation. Our study provides insightful implications for FL practitioners. On the one hand, our findings suggest that FL algorithm designers consider necessary heterogeneity during the evaluation. On the other hand, our findings urge system providers to design specific mechanisms to mitigate the impacts of heterogeneity.

Shuwan Cui, Hao Li, Yanzhao Su et al.

Mandatory lane-changing in complex driving scenarios poses significant challenges for autonomous driving systems due to complex vehicle interactions and strict safety requirements. Existing methods often rely on handcrafted rules or extensive expert demonstrations, which increase data collection costs and provide limited safety guarantees during learning. To address these issues, this paper proposes a rule-guided reinforcement learning framework for lane-changing policy optimization. A lightweight rule-based controller is employed to generate initial experience, guiding the training of an improved Distributional Soft Actor–Critic with Three Refinements (DSAC-T), while a safety-aware constraint controller filters high-risk actions to ensure stable and safe learning. The proposed method is evaluated in Regular Lane Change and Lane Merging scenarios under mixed traffic composed of aggressive and conservative vehicles within a simulation environment. Simulation results show that although lane-changing success rates decrease as traffic aggressiveness increases, the proposed method consistently outperforms SAC and TD3. Notably, under highly aggressive traffic conditions with an aggressiveness ratio of 0.7, the proposed approach improves the success rate by 17.13% compared to SAC and by 10.49% compared to TD3, demonstrating superior robustness and safety in complex, high-conflict lane-changing scenarios. The present study is conducted solely in simulation and requires further validation before application to real-world traffic environments.

Luis S. Piloto, A. Weinstein, P. Battaglia et al.

‘Intuitive physics’ enables our pragmatic engagement with the physical world and forms a key component of ‘common sense’ aspects of thought. Current artificial intelligence systems pale in their understanding of intuitive physics, in comparison to even very young children. Here we address this gap between humans and machines by drawing on the field of developmental psychology. First, we introduce and open-source a machine-learning dataset designed to evaluate conceptual understanding of intuitive physics, adopting the violation-of-expectation (VoE) paradigm from developmental psychology. Second, we build a deep-learning system that learns intuitive physics directly from visual data, inspired by studies of visual cognition in children. We demonstrate that our model can learn a diverse set of physical concepts, which depends critically on object-level representations, consistent with findings from developmental psychology. We consider the implications of these results both for AI and for research on human cognition. Piloto et al. introduce a deep-learning system which is able to learn basic rules of the physical world, such as object solidity and persistence.

Melanie Mcgrath, Andreas Duenser, J. Lacey et al.

Collaborative human-AI (HAI) teaming combines the unique skills and capabilities of humans and machines in sustained teaming interactions leveraging the strengths of each. In tasks involving regular exposure to novelty and uncertainty, collaboration between adaptive, creative humans and powerful, precise artificial intelligence (AI) promises new solutions and efficiencies. User trust is essential to creating and maintaining these collaborative relationships. Established models of trust in traditional forms of AI typically recognize the contribution of three primary categories of trust antecedents: characteristics of the human user, characteristics of the technology, and environmental factors. The emergence of HAI teams, however, requires an understanding of human trust that accounts for the specificity of task contexts and goals, integrates processes of interaction, and captures how trust evolves in a teaming environment over time. Drawing on both the psychological and computer science literature, the process framework of trust in collaborative HAI teams (CHAI-T) presented in this paper adopts the tripartite structure of antecedents established by earlier models, while incorporating team processes and performance phases to capture the dynamism inherent to trust in teaming contexts. These features enable active management of trust in collaborative AI systems, with practical implications for the design and deployment of collaborative HAI teams.

Gizem Izmir Tunahan, Goksu Tuysuzoglu, Hector Altamirano

This study presents a comprehensive, data-driven investigation into students’ seating preferences within academic library environments, aiming to inform user-centered spatial design. Drawing on over 1.3 million ten-minute passive infrared (PIR) sensor observations collected throughout 2023 at the UCL Bartlett Library, we modeled seat-level occupancy using 24 spatial, environmental, and temporal features through advanced machine learning algorithms. Among the models tested, Categorical Boosting (CatBoost) demonstrated the highest predictive performance, achieving a classification accuracy of 72.5%, with interpretability enhanced through SHAP (Shapley Additive exPlanations) analysis. Findings reveal that seating behavior is shaped not by individual factors but by two dominant dimensions: (1) environmental controllability, including access to personal lighting and fresh air, and (2) distraction management, characterized by quiet surroundings, visual privacy, and low-stimulation workspace finishes. In contrast, features commonly presumed to be influential, such as desk width, fixed computer availability, or daylight alone, had minimal impact on seat choice. Despite extensive modeling and optimization, prediction accuracy plateaued at approximately 72%, reflecting the complexity and variability of human behavior in shared learning environments. By integrating long-term behavioral data with explainable machine learning, this study advances the evidence base for academic library design and offers actionable insights. These findings support design strategies that prioritize individual environmental control, as well as acoustic and visual privacy, offering actionable, evidence-based guidance for creating academic library environments that better support student comfort, focus, and engagement.

Jingxuan Wu, Tomoki Saizaki, Tatsuo Yoshinobu et al.

The monitoring of ion homeostasis in vivo is of paramount importance due to its critical functions in biological systems. However, current leading technologies for creating ion-selective electrodes often fall short of the requirements for in vivo applications in terms of multiplexity, miniaturization, and flexibility. To address this gap, we introduce an integrated multiplexed ion monitoring probe created from thermally drawn multi-electrode polymer fiber, aimed at enhancing in vivo ion homeostasis studies. This probe employs a carbon nanofiber (CNF)/graphene composite as the sensing material, utilizing a thermal drawing process, laser machining, and material functionalization to fabricate multiplexed ion probes. Our design incorporates electrodes on micron-scale fibers for sensing Na+, K+ and Cl- ions, alongside an electrode for electrophysiology recording, achieving excellent sensitivity, stability, selectivity, and reversibility in distilled water and artificial cerebrospinal fluid solutions (aCSF). These results demonstrate the potential of the probe for future in vivo applications.

D. Stephen, P. Sethuramalingam

Junya Tang, Y. Liu, Kuo-Yi Lin et al.

Drawing the strengths of data science and machine learning, process mining has recently emerged as an effective research approach for process management and its decision support. Bottleneck identification and analysis is a key problem in process mining which is considered a critical component for process improvement. While previous studies focusing on bottlenecks have been reported, visible gaps remain. Most of these studies considered bottleneck identification from local perspectives by quantitative metrics, such as machine operation and resource requirement, which can not be applied to information and knowledge-intensive processes. Moreover, the root cause of such bottlenecks has not been given enough attention, which limits the impact of process optimisation. This paper proposes an approach that utilises fusion-based clustering and hyperbolic neural network-based knowledge graph embedding for bottleneck identification and root cause analysis. Firstly, a fusion-based clustering is proposed to identify bottlenecks automatically from a global perspective, where the execution frequency of each stage at different periods is calculated to reveal the abnormal stage. Secondly, a process knowledge graph representing tasks, organisations, workforce and relation features as hierarchical and logical patterns is established. Finally, a hyperbolic cluster-based community detection mechanism is researched, based on the process knowledge graph embedding trained by a hyperbolic neural network, to analyse the root cause from a process perspective. Experimental studies using real-world data collected from a multidisciplinary design project revealed the merits of the proposed approach. The execution of the proposed approach is not limited to event logs; it can automatically identify bottlenecks without local quantitative metrics and analyse the causes from a process perspective.

Marvin Xhemrishi, Johan Östman, A. Wachter-Zeh et al.

Federated learning (FL) has emerged as a promising approach for collaboratively training machine learning models while preserving data privacy. Due to its decentralized nature, FL is vulnerable to poisoning attacks, where malicious clients compromise the global model through altered data or updates. Identifying such malicious clients is crucial for ensuring the integrity of FL systems. This task becomes particularly challenging under privacy-enhancing protocols such as secure aggregation, creating a fundamental trade-off between privacy and security. In this work, we propose FedGT, a novel framework designed to identify malicious clients in FL with secure aggregation while preserving privacy. Drawing inspiration from group testing, FedGT leverages overlapping groups of clients to identify the presence of malicious clients via a decoding operation. The clients identified as malicious are then removed from the model training, which is performed over the remaining clients. By choosing the size, number, and overlap between groups, FedGT strikes a balance between privacy and security. Specifically, the server learns the aggregated model of the clients in each group—vanilla federated learning and secure aggregation correspond to the extreme cases of FedGT with group size equal to one and the total number of clients, respectively. The effectiveness of FedGT is demonstrated through extensive experiments on three datasets in a cross-silo setting under different data-poisoning attacks. These experiments showcase FedGT’s ability to identify malicious clients, resulting in high model utility. We further show that FedGT significantly outperforms the private robust aggregation approach based on the geometric median recently proposed by Pillutla et al. and the robust aggregation technique Multi-Krum in multiple settings.

Ernst Tomasch, Heinz Hoschopf, Karin Ausserer et al.

Accidents involving cyclists and trucks are among the most severe road accidents. In 2021, 199 cyclists were killed in accidents involving a truck in the EU. The main accident situation is a truck turning right and a cyclist going straight ahead. A large proportion of these accidents are caused by the inadequate visibility in an HGV (Heavy Goods Vehicle). The blind spot, in particular, is a significant contributor to these accidents. A BSD (Blind Spot Detection) system is expected to significantly reduce these accidents. There are only a few studies that estimate the potential of assistance systems, and these studies include a combined assessment of cyclists and pedestrians. In the present study, accident simulations are used to assess a warning and an autonomously intervening assistance system that could prevent truck to cyclist accidents. The main challenges are local sight obstructions such as fences, hedges, etc., rule violations by cyclists, and the complexity of correctly predicting the cyclist’s intentions, i.e., detecting the trajectory. Taking these accident circumstances into consideration, a BSD system could prevent between 26.3% and 65.8% of accidents involving HGVs and cyclists.

Čuchor Dávid, Bronček Jozef, Obertová Veronika et al.

This study presents an investigation into the tribological, corrosion, and tribocorrosion properties of AISI 316Ti (austenitic) and AISI 430 (ferritic) stainless steels. The comparative analysis focuses on microstructural characterization, hardness, and a series of tribological, electrochemical, and tribocorrosion tests conducted in 0.9% NaCl using a specialized linear tribometer to reveal the quality of the studied materials in tribocorrosion applications. Friction tests were performed under both dry and corrosive conditions, while tribocorrosion tests were conducted under open circuit potential (OCP) conditions in 0.9% NaCl, with the electrode potential of the test specimen monitored during friction. To evaluate the electrochemical behavior of the materials, potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) were conducted using a 0.9% NaCl solution. The measured corrosion potential (Ecorr) suggests that AISI 430 is thermodynamically more stable than AISI 316Ti; however, AISI 316Ti demonstrated higher polarization resistance (RP) values compared to AISI 430. The findings indicate that material qualities significantly influence the coefficient of friction (CoF). Additionally, a notable antifriction effect of 0.9% NaCl was observed during tribological testing, resulting in a lower CoF compared to dry friction conditions. A cathodic shift in OCP during tribocorrosion testing was also observed in both materials, indicating an increase in corrosion vulnerability when the passive layer is degraded.

Alessio Cascino, Enrico Meli, Andrea Rindi

Rolling stock manufacturers are finding innovative structural solutions to improve the quality and reliability of railway vehicles components. Structural optimization processes represent an effective strategy for reducing manufacturing costs, resulting in geometries easier to design and produce. In this framework, the present paper proposes a new methodology to design a railway metro bogie frame, combining structural–topological optimization methods and sensitivity analysis. In addition, manufacturing constraints were included to make the component design suitable for production through sand-casting. A robust sensitivity analysis has highlighted the most critical load conditions acting on the bogie frame. Its effectiveness was verified by carrying out two different structural optimizations based on different loadings. Two equivalent designs were obtained. Computational times were positively reduced by about 57%. The maximum value of stress was reduced about 23%. This new methodology has shown encouraging results to streamline the design process of this complex mechanical system, allowing researchers to also include manufacturing requirements.

Murat Demiral

Steel T-profiles are extensively used across various sectors due to their versatility and reliability. Spot welding plays a crucial role in their production. These profiles are subjected to cyclic bending loads in numerous engineering applications. Understanding the failure mechanisms is essential for enhancing fatigue resistance and extending the operational lifespan of spot-welded assemblies. Key aspects include accurately predicting where damage initiates, how it propagates under increasing cyclic loads, and the failure point. For this purpose, XFEM analysis was conducted and validated with experimental results from the literature. The study emphasizes the significant impact of bending moment magnitude, load ratio, the diameter of spot welds, and component thickness on the fatigue performance of spot-welded assemblies under bending loads. All these parameters significantly affected the fatigue response. Notably, thinner components showed 8.55 times faster crack propagation, accompanied by more localized and severe cracking.

Akira Okuno, Shunsuke Yoshimoto, Akio Yamamoto

Abstract A built-in capacitive proximity sensing method for a charge-induction electrostatic film actuator is proposed. This actuator consists of two thin sheets that function as a stator and a slider. A stator is an insulating sheet with many strips of electrodes in it, whereas a slider is a dielectric sheet that has slight conductivity on its surface. By applying actuation voltage on stator electrodes, the slider that is placed on the stator is driven by electrostatic force. This research realized the simultaneous actuation and proximity sensing using the same electrodes by integrating a resonance-based capacitance measurement circuit into a driving circuit. The study investigated the impact of having a slider on sensing performance, confirming the feasibility of simultaneous sensing and driving. The implemented system achieved an interactive actuation that changed driving velocity according to the proximity distance of the human hand.

J. Lim, Teemu Leinonen, Lasse Lipponen et al.

ABSTRACT Artificial Intelligence (AI) has significantly advanced in creating professional-level media content. In creative education, determining how students can benefit without becoming dependent on them is a challenge. In this study, researchers conducted an exploratory experiment that positioned AI as a relational artifact to students in a series of drawing activities and examined the potential impact of affective relations with machines in socio-cultural creative learning. The resulting artifacts, observations, and interview transcripts were analyzed using the Consensual Assessment Technique and a grounded theory approach. The study's results indicate that the design professors reliably evaluated the student drawings as more creative than the AI drawings, but neither demonstrated a consistent increase in creativity. However, the presence of AI engaged the students to explore different approaches to artistic prompts. We theorize that AI can be mediated as a learning artifact for transformative creativity if the students perceive their relationship with AI as empathetic and collaborative.

Aleksa Bisercic, Mladen Nikolic, M. Schaar et al.

Tabular data is often hidden in text, particularly in medical diagnostic reports. Traditional machine learning (ML) models designed to work with tabular data, cannot effectively process information in such form. On the other hand, large language models (LLMs) which excel at textual tasks, are probably not the best tool for modeling tabular data. Therefore, we propose a novel, simple, and effective methodology for extracting structured tabular data from textual medical reports, called TEMED-LLM. Drawing upon the reasoning capabilities of LLMs, TEMED-LLM goes beyond traditional extraction techniques, accurately inferring tabular features, even when their names are not explicitly mentioned in the text. This is achieved by combining domain-specific reasoning guidelines with a proposed data validation and reasoning correction feedback loop. By applying interpretable ML models such as decision trees and logistic regression over the extracted and validated data, we obtain end-to-end interpretable predictions. We demonstrate that our approach significantly outperforms state-of-the-art text classification models in medical diagnostics. Given its predictive performance, simplicity, and interpretability, TEMED-LLM underscores the potential of leveraging LLMs to improve the performance and trustworthiness of ML models in medical applications.

Jing Du, William Vann, Tianyu Zhou et al.

In the realm of robotics and automation, robot teleoperation, which facilitates human–machine interaction in distant or hazardous settings, has surged in significance. A persistent issue in this domain is the delays between command issuance and action execution, causing negative repercussions on operator situational awareness, performance, and cognitive load. These delays, particularly in long-distance operations, are difficult to mitigate even with the most advanced computing advancements. Current solutions mainly revolve around machine-based adjustments to combat these delays. However, a notable lacuna remains in harnessing human perceptions for an enhanced subjective teleoperation experience. This paper introduces a novel approach of sensory manipulation for induced human adaptation in delayed teleoperation. Drawing from motor learning and rehabilitation principles, it is posited that strategic sensory manipulation, via altered sensory stimuli, can mitigate the subjective feeling of these delays. The focus is not on introducing new skills or adapting to novel conditions; rather, it leverages prior motor coordination experience in the context of delays. The objective is to reduce the need for extensive training or sophisticated automation designs. A human-centered experiment involving 41 participants was conducted to examine the effects of modified haptic cues in teleoperations with delays. These cues were generated from high-fidelity physics engines using parameters from robot-end sensors or physics engine simulations. The results underscored several benefits, notably the considerable reduction in task time and enhanced user perceptions about visual delays. Real-time haptic feedback, or the anchoring method, emerged as a significant contributor to these benefits, showcasing reduced cognitive load, bolstered self-confidence, and minimized frustration. Beyond the prevalent methods of automation design and training, this research underscores induced human adaptation as a pivotal avenue in robot teleoperation. It seeks to enhance teleoperation efficacy through rapid human adaptation, offering insights beyond just optimizing robotic systems for delay compensations.

Jinming Zheng, Lin Wang, Xiaochan Wang et al.

The discrete element parameters of cabbages (Brassica oleracea L.) were calibrated for the design and parameter optimization of a cabbage harvester. The cabbage model was created based on the study of cabbage material characteristics and the simulation model parameters of cabbage were calibrated. The intrinsic parameters and partial contact parameters of cabbages were obtained by direct measurement. The cabbage accumulation angle was determined by a plate drawing test. Through the steepest ascent test and the orthogonal rotation combination test, a regression model of the cabbage accumulation angle error was established. The optimal contact parameters between the cabbages were obtained by the minimum error modeling. These calibrated parameters were applied in the verification test, and the results indicated that the error between the simulated and measured values of the cabbage accumulation angle was only 1.63%, which demonstrated that the results were dependable. This study can provide a theoretical support for designing and optimizing the parameters of cabbage harvesting machines with the discrete element method (DEM).