Hasil untuk "Machine design and drawing"

Fatemeh Ghorbani, Augustin Hym, Mohammed Elhenawy et al.

Motorcyclists face a disproportionately high risk of severe injury or death compared to other road users, highlighting the need for intelligent rider assistance technologies. This paper presents an initial, modular, and interpretable AI pipeline that generates context-aware safety advice from first-person motorcycle videos with practical inference latency suitable for on-device deployment, framing large language models as interpretable cognitive support agents for motorcycle safety. The system integrates lightweight perception and reasoning components to emulate the function of an Advanced Rider Assistance System (ARAS). Video frames are processed at 1 FPS using Pixtral, a Mistral-based multimodal large language model (MLLM), to produce descriptive scene captions, while YOLOv8 identifies key objects such as vehicles, pedestrians, and road hazards. A Mistral-small language model then fuses this information to generate concise, imperative safety tips. Preliminary evaluations on publicly available motorcycle POV datasets demonstrate promising performance in terms of contextual accuracy, interpretability, and scalability, suggesting potential for real-world deployment in low-resource or embedded environments. The proposed framework offers interpretable, context-aware safety assistance that is particularly valuable for young and newly licensed riders during the transition from supervised training to independent riding, where real-time hazard interpretation support is most needed.

Amir Hossein Sheikhshoaei, Ali Khoshsima, Ahmadreza Salehi et al.

Air pollution continues to be one of the most critical environmental challenges today. Among the various contaminants released into the atmosphere, ammonia (NH3) has been drawing more attention because of its considerable effects on both the environment and human health. Its emissions contribute to the formation of secondary particulate matter, which not only poses significant health risks but also deteriorates environmental quality. Ionic liquids have emerged as promising candidates for capturing NH3, and the ability to reliably predict its solubility in ILs is essential for identifying suitable solvents and optimizing the separation process. This study uses machine learning models (CatBoost, XGBoost, LightGBM, and GPR) for predicting ammonia solubility in ionic liquids. The input parameters include temperature (T), pressure (P), critical temperature (Tc), critical pressure (Pc), critical volume (Vc), acentric factor (ω), and boiling point (Tb). Statistical errors and graphical analyses showed that the CatBoost model performed better than other models and had high reliability for predicting NH3 solubility. Among the evaluated models, CatBoost delivered the most accurate predictions, achieving a root mean square error (RMSE) of 0.0137 and an R² value of 0.9967 for all data. This model effectively captured the influence of key parameters on ammonia solubility. Notably, 97.47 % of the data points fell within the model’s applicability domain, highlighting its strong predictive reliability. These outcomes underscore the capability of the CatBoost algorithm to serve as a robust and efficient approach for estimating NH3 solubility in ionic liquids, offering valuable support for future materials design and separation process optimization.

Clara Fannjiang, Ji Won Park

Algorithms for machine learning-guided design, or design algorithms, use machine learning-based predictions to propose novel objects with desired property values. Given a new design task -- for example, to design novel proteins with high binding affinity to a therapeutic target -- one must choose a design algorithm and specify any hyperparameters and predictive and/or generative models involved. How can these decisions be made such that the resulting designs are successful? This paper proposes a method for design algorithm selection, which aims to select design algorithms that will produce a distribution of design labels satisfying a user-specified success criterion -- for example, that at least ten percent of designs' labels exceed a threshold. It does so by combining designs' predicted property values with held-out labeled data to reliably forecast characteristics of the label distributions produced by different design algorithms, building upon techniques from prediction-powered inference. The method is guaranteed with high probability to return design algorithms that yield successful label distributions (or the null set if none exist), if the density ratios between the design and labeled data distributions are known. We demonstrate the method's effectiveness in simulated protein and RNA design tasks, in settings with either known or estimated density ratios.

Kaiqu Liang, Haimin Hu, Xuandong Zhao et al.

Bullshit, as conceptualized by philosopher Harry Frankfurt, refers to statements made without regard to their truth value. While previous work has explored large language model (LLM) hallucination and sycophancy, we propose machine bullshit as an overarching conceptual framework that can allow researchers to characterize the broader phenomenon of emergent loss of truthfulness in LLMs and shed light on its underlying mechanisms. We introduce the Bullshit Index, a novel metric quantifying LLMs' indifference to truth, and propose a complementary taxonomy analyzing four qualitative forms of bullshit: empty rhetoric, paltering, weasel words, and unverified claims. We conduct empirical evaluations on the Marketplace dataset, the Political Neutrality dataset, and our new BullshitEval benchmark (2,400 scenarios spanning 100 AI assistants) explicitly designed to evaluate machine bullshit. Our results demonstrate that model fine-tuning with reinforcement learning from human feedback (RLHF) significantly exacerbates bullshit and inference-time chain-of-thought (CoT) prompting notably amplify specific bullshit forms, particularly empty rhetoric and paltering. We also observe prevalent machine bullshit in political contexts, with weasel words as the dominant strategy. Our findings highlight systematic challenges in AI alignment and provide new insights toward more truthful LLM behavior.

Bikram Khanal, Pablo Rivas, Arun Sanjel et al.

Despite the mounting anticipation for the quantum revolution, the success of Quantum Machine Learning (QML) in the Noisy Intermediate-Scale Quantum (NISQ) era hinges on a largely unexplored factor: the generalization error bound, a cornerstone of robust and reliable machine learning models. Current QML research, while exploring novel algorithms and applications extensively, is predominantly situated in the context of noise-free, ideal quantum computers. However, Quantum Circuit (QC) operations in NISQ-era devices are susceptible to various noise sources and errors. In this article, we conduct a Systematic Mapping Study (SMS) to explore the state-of-the-art generalization bound for supervised QML in NISQ-era and analyze the latest practices in the field. Our study systematically summarizes the existing computational platforms with quantum hardware, datasets, optimization techniques, and the common properties of the bounds found in the literature. We further present the performance accuracy of various approaches in classical benchmark datasets like the MNIST and IRIS datasets. The SMS also highlights the limitations and challenges in QML in the NISQ era and discusses future research directions to advance the field. Using a detailed Boolean operators query in five reliable indexers, we collected 544 papers and filtered them to a small set of 37 relevant articles. This filtration was done following the best practice of SMS with well-defined research questions and inclusion and exclusion criteria.

Ralf Stetter

Automated guided vehicles undertake complex transportation tasks, for instance, in production and storage systems. In recent years, an increased focus on sustainability has occurred as the effects of ongoing climate change have become more apparent. Engineers are searching intensively for ways to design technical systems that are not only environmentally sustainable, but are also resilient to the challenges of the changing climate and other environmental conditions. The production of automated guided vehicles requires considerable resources; therefore, a long operation time is desirable for overall sustainability. The performance of transportation tasks requires certain processes, such as control, path planning, coordination/synchronization, and maintenance and update processes—the latter are also very important for a long operation time. This article proposes understanding these processes as services and to explore product service systems with automated guided vehicles. Due to their complexity, the efficient and safe operation of such systems can be at risk because of several factors, such as component faults, external attacks and disturbances. For several years both resilient control and resilience engineering have been researched as possible remedies. An extension of these two concepts to the early stages of system development processes and including the system’s hardware is proposed in this article. This extension is referred to as resilient design. A primary purpose of resilient design is sustainability through extended usability and planned updates. The main intention of this article is to provide a comprehensive understanding of resilient design through application to product service systems with automated guided vehicles. The basis for this contribution is an extensive literature review and detailed system analyses on different levels. The main research results include novel application modes for product development methods. The explanation of the results is supported by means of an illustrative example based on a product service system with automated guided vehicles.

Ramanjeet Singh, Jing Ren, Xianke Lin

Path planning is the most fundamental necessity for autonomous mobile robots. Traditionally, the path planning problem was solved using analytical methods, but these methods need perfect localization in the environment, a fully developed map to plan the path, and cannot deal with complex environments and emergencies. Recently, deep neural networks have been applied to solve this complex problem. This review paper discusses path-planning methods that use neural networks, including deep reinforcement learning, and its different types, such as model-free and model-based, <i>Q</i>-value function-based, policy-based, and actor-critic-based methods. Additionally, a dedicated section delves into the nuances and methods of robot interactions with pedestrians, exploring these dynamics in diverse environments such as sidewalks, road crossings, and indoor spaces, underscoring the importance of social compliance in robot navigation. In the end, the common challenges faced by these methods and applied solutions such as reward shaping, transfer learning, parallel simulations, etc. to optimize the solutions are discussed.

Ashwin Ram

This paper represents a groundbreaking advancement in Parkinson disease (PD) research by employing a novel machine learning framework to categorize PD into distinct subtypes and predict its progression. Utilizing a comprehensive dataset encompassing both clinical and neurological parameters, the research applies advanced supervised and unsupervised learning techniques. This innovative approach enables the identification of subtle, yet critical, patterns in PD manifestation, which traditional methodologies often miss. Significantly, this research offers a path toward personalized treatment strategies, marking a major stride in the precision medicine domain and showcasing the transformative potential of integrating machine learning into medical research.

Puguh Elmiawan, Dharmanto Dharmanto, Adik S.W et al.

This paper discusses the building of a low-cost CNC router machine. In designing a CNC machine using an electrical actuator as the driving force, we use a nema 23 stepper motor combined with a pitch 2 mm timing belt to drive the machine on the x and y axes. On the z-axis, we use a stepper motor and a linear screw to drive the machine. This machine is often used to cut and engrave wood, acrylic, and PCB objects. The design drawing that has been made on the PC is sent to the microcontroller using serial communication, then the CNC performs work on the thing according to the coordinates. The spindle will create a pattern on the thing automatically according to the design drawing. Measurement was made using a caliper with an accuracy of 0.05 mm.  This CNC machine is performing duly as ordered by the mach3 micro controller. The machine test results show an accuracy value of 99.85% on the x-axis, 99.74% on the y-axis, and 99.12% on the z-axis.

Daniel Furelos-Blanco, Mark Law, Anders Jonsson et al.

Reward machines (RMs) are a recent formalism for representing the reward function of a reinforcement learning task through a finite-state machine whose edges encode subgoals of the task using high-level events. The structure of RMs enables the decomposition of a task into simpler and independently solvable subtasks that help tackle long-horizon and/or sparse reward tasks. We propose a formalism for further abstracting the subtask structure by endowing an RM with the ability to call other RMs, thus composing a hierarchy of RMs (HRM). We exploit HRMs by treating each call to an RM as an independently solvable subtask using the options framework, and describe a curriculum-based method to learn HRMs from traces observed by the agent. Our experiments reveal that exploiting a handcrafted HRM leads to faster convergence than with a flat HRM, and that learning an HRM is feasible in cases where its equivalent flat representation is not.

Alicia Curth, Alihan Hüyük, Mihaela van der Schaar

We study the problem of adaptively identifying patient subpopulations that benefit from a given treatment during a confirmatory clinical trial. This type of adaptive clinical trial has been thoroughly studied in biostatistics, but has been allowed only limited adaptivity so far. Here, we aim to relax classical restrictions on such designs and investigate how to incorporate ideas from the recent machine learning literature on adaptive and online experimentation to make trials more flexible and efficient. We find that the unique characteristics of the subpopulation selection problem -- most importantly that (i) one is usually interested in finding subpopulations with any treatment benefit (and not necessarily the single subgroup with largest effect) given a limited budget and that (ii) effectiveness only has to be demonstrated across the subpopulation on average -- give rise to interesting challenges and new desiderata when designing algorithmic solutions. Building on these findings, we propose AdaGGI and AdaGCPI, two meta-algorithms for subpopulation construction. We empirically investigate their performance across a range of simulation scenarios and derive insights into their (dis)advantages across different settings.

L. Peltier, O. Perroud, P. Moll et al.

Hossam A. Gabbar, Yasser Elsayed, Abu Bakar Siddique et al.

The popularity of the eBus has been increasing rapidly in recent years due to its low greenhouse gases (GHG) emissions and its low dependence on fossil fuels. This incremental use of the eBus increases the burden to the power grid for its charging. Charging eBus requires a high amount of power for a feasible amount of time. Therefore, developing a fast-charging station (FCS) integrated with Micro Energy Grid (MEG) and hybrid energy storage is crucial for charging eBuses. This paper presents a design of FCS for eBus that integrates MEG with hybrid energy storage with the energy management system. To reduce the dependency on the main utility grid, a hybrid micro energy grid based on a renewable source (i.e., PV) have been included. In addition, hybrid energy storage of batteries and flywheels has also been developed to mitigate the power demand of the fast-charging station during peak time. Furthermore, a multiple-input DC-DC converter has been developed for managing the DC power transfer between the common DC bus and the multiple energy sources. Finally, an energy management system and the controller has been designed to achieve an extensive performance from the fast charging station. MATLAB Simulink has been used for the simulation work of the overall design. Different test case scenarios are tested for evaluating the performance parameters of the proposed FCS and also for evaluating its performance.

Pattanapairoj Sirorat, Nitisiri Krisanarach, Sethanan Kanchana

Industry 4.0 is an era in which the manufacturing industry has adopted digital technologies and the Internet to enable smart manufacturing system, machines used in the production now can communicate with each other and exchange information between each other, and the machinery used in the manufacturing process is more modern and precise. Therefore, educational institutions should develop the curriculum to produce qualified graduates with the knowledge required for the Industry 4.0 era, especially Industrial Engineering graduates who are directly related to the industry sector. The purpose of this research is to collect the data for the Master of Industrial Engineering (MSIE) curriculum development. The Analytic Hierarchy Process (AHP) technique is used to rank the indicators of knowledge that is important to the employment of graduates with a master’s degree in Industrial Engineering, and study the gap between the expectations of employers and the ability of the current MSIE students of Khon Kaen University. The results of the study reveal that the first indicators that are most important to the employment of MSIE graduates is the knowledge of Industry 4.0 strategy and the knowledge that the students should have developed are the collaboration of humans and robots, big data analytics, real time data usage and databased decision making.

Sousuke Nakamura, Shunsuke Muto, Daichi Takahashi

Abstract Simultaneous localization and mapping (SLAM) is a widely used technology in autonomous mobile robots, where sensors such as Lidar or cameras are typically used. Sensor fusion using multiple sensors has been employed to compensate for the shortcomings of each sensor in SLAM. However, the sensor cost cannot be ignored when considering its practical usage. Therefore, this study aims at realizing a high-precision SLAM using a sensor switching system, combining multiple low-cost sensors. The sensor switching system consists of a low-cost Lidar SLAM and a monocular localization. Since a low-cost Lidar has a short laser range, degeneracy often occurs due to the fact that they cannot capture features while building maps. The proposed system uses localization data from monocular localization to ensure precision in regions where degeneracy occurs. The proposed system was evaluated through the simulation assuming the museum environment where the degeneracy occurred. The accuracy of the robot trajectory and the built map proved the effectiveness of the proposed system.

Hui Zhang, Rongqing Zhang, Chen Chen et al.

In this paper, we investigate the intersection traffic management for connected automated vehicles (CAVs). In particular, a decentralized autonomous intersection management scheme that takes into account both the traffic efficiency and scheduling flexibility is proposed, which adopts a novel intersection–vehicle model to check conflicts among CAVs in the entire intersection area. In addition, a priority-based collision-avoidance rule is set to improve the performance of traffic efficiency and shorten the delays of emergency CAVs. Moreover, a multi-objective function is designed to obtain the optimal trajectories of CAVs, which considers ride comfort, velocities of CAVs, fuel consumption, and the constraints of safety, velocity, and acceleration. Simulation results demonstrate that our proposed scheme can achieve good performance in terms of traffic efficiency and shortening the delays of emergency CAVs.

Ane Dalsnes Storsæter, Kelly Pitera, Edward McCormack

Road markings are beneficial to human drivers, advanced driver assistance systems (ADAS), and automated driving systems (ADS); on the contrary, snow coverage on roads poses a challenge to all three of these groups with respect to lane detection, as white road markings are difficult to distinguish from snow. Indeed, yellow road markings provide a visual contrast to snow that can increase a human drivers’ visibility. Yet, in spite of this fact, yellow road markings are becoming increasingly rare in Europe due to the high costs of painting and maintaining two road marking colors. More importantly, in conjunction with our increased reliance on automated driving, the question of whether yellow road markings are of value to automatic lane detection functions arises. To answer this question, images from snowy conditions are assessed to see how different representations of colors in images (color spaces) affect the visibility levels of white and yellow road markings. The results presented in this paper suggest that yellow markings provide a certain number of benefits for automated driving, offering recommendations as to what the most appropriate color spaces are for detecting lanes in snowy conditions. To obtain the safest and most cost-efficient roads in the future, both human and automated drivers’ actions must be considered. Road authorities and car manufacturers also have a shared interest in discovering how road infrastructure design, including road marking, can be adapted to support automated driving.

Qihao Zhu, Jianxi Luo

Novel concepts are essential for design innovation and can be generated with the aid of data stimuli and computers. However, current generative design algorithms focus on diagrammatic or spatial concepts that are either too abstract to understand or too detailed for early phase design exploration. This paper explores the uses of generative pre-trained transformers (GPT) for natural language design concept generation. Our experiments involve the use of GPT-2 and GPT-3 for different creative reasonings in design tasks. Both show reasonably good performance for verbal design concept generation.

Yves-Laurent Kom Samo

We introduce the first application of the lean methodology to machine learning projects. Similar to lean startups and lean manufacturing, we argue that lean machine learning (LeanML) can drastically slash avoidable wastes in commercial machine learning projects, reduce the business risk in investing in machine learning capabilities and, in so doing, further democratize access to machine learning. The lean design pattern we propose in this paper is based on two realizations. First, it is possible to estimate the best performance one may achieve when predicting an outcome $y \in \mathcal{Y}$ using a given set of explanatory variables $x \in \mathcal{X}$, for a wide range of performance metrics, and without training any predictive model. Second, doing so is considerably easier, faster, and cheaper than learning the best predictive model. We derive formulae expressing the best $R^2$, MSE, classification accuracy, and log-likelihood per observation achievable when using $x$ to predict $y$ as a function of the mutual information $I\left(y; x\right)$, and possibly a measure of the variability of $y$ (e.g. its Shannon entropy in the case of classification accuracy, and its variance in the case regression MSE). We illustrate the efficacy of the LeanML design pattern on a wide range of regression and classification problems, synthetic and real-life.

Cong Hao, Jordan Dotzel, Jinjun Xiong et al.

Artificial intelligence (AI) technologies have dramatically advanced in recent years, resulting in revolutionary changes in people's lives. Empowered by edge computing, AI workloads are migrating from centralized cloud architectures to distributed edge systems, introducing a new paradigm called edge AI. While edge AI has the promise of bringing significant increases in autonomy and intelligence into everyday lives through common edge devices, it also raises new challenges, especially for the development of its algorithms and the deployment of its services, which call for novel design methodologies catered to these unique challenges. In this paper, we provide a comprehensive survey of the latest enabling design methodologies that span the entire edge AI development stack. We suggest that the key methodologies for effective edge AI development are single-layer specialization and cross-layer co-design. We discuss representative methodologies in each category in detail, including on-device training methods, specialized software design, dedicated hardware design, benchmarking and design automation, software/hardware co-design, software/compiler co-design, and compiler/hardware co-design. Moreover, we attempt to reveal hidden cross-layer design opportunities that can further boost the solution quality of future edge AI and provide insights into future directions and emerging areas that require increased research focus.