Search Results (15)

Deep neural network (DNN) hardware (HW) accelerators have achieved great success in improving DNNs’ performance and efficiency. One key reason is the dataflow in executing a DNN layer, including on-chip data partitioning, computation parallelism, and scheduling policy, which have large impacts on latency and energy consumption. Unlike prior works that required considerable efforts from HW engineers to design suitable dataflows for different DNNs, this work proposes an efficient data-centric approach, named Dataflow Code Propagation (DCP), to automatically find the optimal dataflow for DNN layers in seconds without human effort. It has several attractive benefits that prior studies lack, including the following: (i) We translate the HW dataflow configuration into a code representation in a unified dataflow coding space, which can be optimized by back-propagating gradients given a DNN layer or network. (ii) DCP learns a neural predictor to efficiently update the dataflow codes towards the desired gradient directions to minimize various optimization objectives, e.g., latency and energy. (iii) It can be easily generalized to unseen HW configurations in a zero-shot or few-shot learning manner. For example, without using additional training data, Extensive experiments on several representative models such as MobileNet, ResNet, and ViT show that DCP outperforms its counterparts in various settings. Full article

The On-Board Diagnostics II (OBD-II) system, driven by a wide range of embedded sensors, has revolutionized the automotive industry by enabling real-time monitoring of key vehicle parameters such as engine load, vehicle speed, throttle position, and diagnostic trouble codes. Concurrently, recent advancements in machine learning (ML) have further expanded the capabilities of OBD-II applications, unlocking advanced, intelligent, and data-centric functionalities that significantly surpass those of conventional methodologies. This paper presents a comprehensive investigation into ML-based applications that leverage OBD-II sensor data, aiming to enhance sustainability, operational efficiency, safety, and security in modern vehicular systems. To this end, a diverse set of ML approaches is examined, encompassing supervised, unsupervised, reinforcement learning (RL), deep learning (DL), and hybrid models intended to support advanced driving analytics tasks such as fuel optimization, emission control, driver behavior analysis, anomaly detection, cybersecurity, road perception, and driving support. Furthermore, this paper synthesizes recent research contributions and practical implementations, identifies prevailing challenges, and outlines prospective research directions. Full article

This paper presents a novel application of machine learning models to clarify the intricate behaviors of expansive soils, focusing on the impact of sand content, saturation level, and dry density. Departing from conventional methods, this research utilizes a data-centric approach, employing a suite of sophisticated machine learning models to predict soil properties with remarkable precision. The inclusion of a 30% sand mixture is identified as a critical threshold for optimizing soil strength and stiffness, a finding that underscores the transformative potential of sand amendment in soil engineering. In a significant advancement, the study benchmarks the predictive power of several models including extreme gradient boosting (XGBoost), gradient boosting regression (GBR), random forest regression (RFR), decision tree regression (DTR), support vector regression (SVR), symbolic regression (SR), and artificial neural networks (ANNs and proposed ANN-GMDH). Symbolic regression equations have been developed to predict the elasticity modulus and unconfined compressive strength of the investigated expansive soil. Despite the complex behaviors of expansive soil, the trained models allow for optimally predicting the values of unconfined compressive parameters. As a result, this paper provides for the first time a reliable and simply applicable approach for estimating the unconfined compressive parameters of expansive soils. The proposed ANN-GMDH model emerges as the pre-eminent model, demonstrating exceptional accuracy with the best metrics. These results not only highlight the ANN’s superior performance but also mark this study as a groundbreaking endeavor in the application of machine learning to soil behavior prediction, setting a new benchmark in the field. Full article

Speed reducers (SR) and electric motors are crucial in modern manufacturing, especially within adhesive coating equipment. The electric motor mainly transforms electrical power into mechanical force to propel most machinery. Conversely, speed reducers are vital elements that control the speed and torque of rotating machinery, ensuring optimal performance and efficiency. Interestingly, variations in chamber temperatures of adhesive coating machines and the use of specific adhesives can lead to defects in chains and jigs, causing possible breakdowns in the speed reducer and its surrounding components. This study introduces novel deep-learning autoencoder models to enhance production efficiency by presenting a comparative assessment for anomaly detection that would enable precise and predictive insights by modeling complex temporal relationships in the vibration data. The data acquisition framework facilitated adherence to data governance principles by maintaining data quality and consistency, data storage and processing operations, and aligning with data management standards. The study here would capture the attention of practitioners involved in data-centric processes, industrial engineering, and advanced manufacturing techniques. Full article

Metal processing may benefit from innovative lean-and-green datacentric engineering techniques. Broad process improvement opportunities in the efficient usage of materials and energy are anticipated (United Nations Sustainable Development Goals #9, 12). A CO₂ laser cutting method is investigated in this study in terms of product characteristics (surface roughness (SR)) and process characteristics (energy (EC) and gas consumption (GC) as well as cutting time (CT)). The examined laser cutter controlling factors were as follows: (1) the laser power (LP), (2) the cutting speed (CS), (3) the gas pressure (GP) and, (4) the laser focus length (F). The selected 10mm-thick carbon steel (EN10025 St37-2) workpiece was arranged to have various geometric configurations so as to simulate a variety of real industrial milling demands. Non-linear saturated screening/optimization trials were planned using the Taguchi-type L₉(3⁴) orthogonal array. The resulting multivariate dataset was treated using a combination of the Gibbs sampler and the Pareto frontier method in order to approximate the strength of the studied effects and to find a solution that comprises the minimization of all the tested process/product characteristics. The Pareto frontier optimal solution was (EC, GC, CT, SR) = (4.67 kWh, 20.35 Nm³, 21 s, 5.992 μm) for the synchronous screening/optimization of the four characteristics. The respective factorial settings were optimally adjusted at the four inputs (LP, CS, GP, F) located at (4 kW, 1.9 mm/min, 0.75 bar, +2.25 mm). The linear regression analysis was aided by the Gibbs sampler and promoted the laser power and the cutting speed on energy consumption to be stronger effects. Similarly, a strong effect was identified of the cutting speed and the gas pressure on gas consumption as well as a reciprocal effect of the cutting speed on the cutting time. Further industrial explorations may involve more intricate workpiece geometries, burr formation phenomena, and process economics. Full article

Collaboration among individuals with diverse skills and personalities is crucial to producing high-quality software. The success of any software project depends on the team’s cohesive functionality and mutual complementation. This study introduces a data-centric methodology for forming Software Engineering (SE) teams centred around personality traits. Our study analysed data from an SE course where 157 students in 31 teams worked through four project phases and were evaluated based on deliverables and instructor feedback. Using the Five-Factor Model (FFM) and a variety of statistical tests, we determined that teams with higher levels of extraversion and conscientiousness, and lower neuroticism, consistently performed better. We examined team members’ interactions and developed a predictive model using extreme gradient boosting. The model achieved a 74% accuracy rate in predicting inter-member satisfaction rankings. Through graphical explainability, the model underscored incompatibilities among members, notably those with differing levels of extraversion. Based on our findings, we introduce a team formation algorithm using Simulated Annealing (SA) built upon the insights derived from our predictive model and additional heuristics. Full article

Water scarcity is a challenging global risk. Urban wastewater treatment technologies, which utilize processes based on single-stage ultrafiltration (UF) or nanofiltration (NF), have the potential to offer lean-and-green cost-effective solutions. Robustifying the effectiveness of water treatment is a complex multidimensional characteristic problem. In this study, a non-linear Taguchi-type orthogonal-array (OA) sampler is enriched with an emergent stigmergic clustering procedure to conduct the screening/optimization of multiple UF/NF aquametric performance metrics. The stochastic solver employs the Databionic swarm intelligence routine to classify the resulting multi-response dataset. Next, a cluster separation measure, the Davies–Bouldin index, is used to evaluate input and output relationships. The self-organized bionic-classifier data-partition appropriateness is matched for signatures between the emergent stigmergic clustering memberships and the OA factorial vector sequences. To illustrate the proposed methodology, recently-published multi-response multifactorial L₉(3⁴) OA-planned experiments from two interesting UF-/NF-membrane processes are examined. In the study, seven UF-membrane process characteristics and six NF-membrane process characteristics are tested (1) in relationship to four controlling factors and (2) to synchronously evaluate individual factorial curvatures. The results are compared with other ordinary clustering methods and their performances are discussed. The unsupervised robust bionic prediction reveals that the permeate flux influences both the UF-/NF-membrane process performances. For the UF process and a three-cluster model, the Davies–Bouldin index was minimized at values of 1.89 and 1.27 for the centroid and medoid centrotypes, respectively. For the NF process and a two-cluster model, the Davies–Bouldin index was minimized for both centrotypes at values close to 0.4, which was fairly close to the self-validation value. The advantage of this proposed data-centric engineering scheme relies on its emergent and self-organized clustering capability, which retraces its appropriateness to the fractional factorial rigid structure and, hence, it may become useful for screening and optimizing small-data wastewater operating conditions. Full article

This work presents a data-centric method to use IoT data, generated from the site, to monitor core functions of safety barriers on a batch reactor. The approach turns process safety performance indicators (PSPIs) into online, globally available safety indicators that eliminate variability in human interpretation. This work also showcases a class of PSPIs that are reliable and time-dependent but only work in a digital online environment: profile PSPIs. It is demonstrated that the profile PSPI opens many new opportunities for leading indicators, without the need for complex mathematics. Online PSPI analyses were performed at the Syngenta Huddersfield Manufacturing Centre, Leeds Road, West Yorkshire, United Kingdom, and shared with their international headquarters in Basel, Switzerland. The performance was determined with industry software to extract time-series data and perform the calculations. The calculations were based on decades of IoT data stored in the AVEVA Factory Historian. Non-trivial data cleansing and additional data tags were required for the creation of relevant signal conditions and composite conditions. This work demonstrates that digital methods do not require gifted data analysts to report existing PSPIs in near real-time and is well within the capabilities of chemical (safety) engineers. Current PSPIs can also be evaluated in terms of their effectiveness to allow management to make decisions that lead to corrective actions. This improves significantly on traditional PSPI processes that, when reviewed monthly, lead to untimely decisions and actions. This approach also makes it possible to review PSPIs as they develop, receiving notifications of PSPIs when they reach prescribed limits, all with the potential to recommend alternative PSPIs that are more proactive in nature. Full article

In recent years, the field of artificial intelligence has experienced significant growth, which has been primarily attributed to advancements in hardware and the efficient training of deep neural networks on graphics processing units. The development of high-quality artificial intelligence solutions necessitates a strong emphasis on data-centric approaches that involve the collection, labeling and quality-assurance of data and labels. These processes, however, are labor-intensive and often demand extensive human effort. Simultaneously, there exists an abundance of untapped data that could potentially be utilized to train models capable of addressing complex problems. These raw data, nevertheless, require refinement to become suitable for machine learning training. This study concentrates on the computer vision subdomain within artificial intelligence and explores data requirements within the context of requirements engineering. Among the various data requirement activities, label quality assurance is crucial. To address this problem, we propose a machine learning-based method for automatic label quality assurance, especially in the context of object detection use cases. Our approach aims to support both annotators and computer vision project stakeholders while reducing the time and resources needed to conduct label quality assurance activities. In our experiments, we trained a neural network on a small set of labeled data and achieved an accuracy of 82% in differentiating good and bad labels on a large set of labeled data. This demonstrates the potential of our approach in automating label quality assurance. Full article

Machine learning (ML) techniques discover knowledge from large amounts of data. Modeling in ML is becoming essential to software systems in practice. The accuracy and efficiency of ML models have been focused on ML research communities, while there is less attention on validating the qualities of ML models. Validating ML applications is a challenging and time-consuming process for developers since prediction accuracy heavily relies on generated models. ML applications are written by relatively more data-driven programming based on the black box of ML frameworks. All of the datasets and the ML application need to be individually investigated. Thus, the ML validation tasks take a lot of time and effort. To address this limitation, we present a novel quality validation technique that increases the reliability for ML models and applications, called MLVal. Our approach helps developers inspect the training data and the generated features for the ML model. A data validation technique is important and beneficial to software quality since the quality of the input data affects speed and accuracy for training and inference. Inspired by software debugging/validation for reproducing the potential reported bugs, MLVal takes as input an ML application and its training datasets to build the ML models, helping ML application developers easily reproduce and understand anomalies in the ML application. We have implemented an Eclipse plugin for MLVal that allows developers to validate the prediction behavior of their ML applications, the ML model, and the training data on the Eclipse IDE. In our evaluation, we used 23,500 documents in the bioengineering research domain. We assessed the ability of the MLVal validation technique to effectively help ML application developers: (1) investigate the connection between the produced features and the labels in the training model, and (2) detect errors early to secure the quality of models from better data. Our approach reduces the cost of engineering efforts to validate problems, improving data-centric workflows of the ML application development. Full article

The idea of a digital twin has recently gained widespread attention. While, so far, it has been used predominantly for problems in engineering and manufacturing, it is believed that a digital twin also holds great promise for applications in medicine and health. However, a problem that severely hampers progress in these fields is the lack of a solid definition of the concept behind a digital twin that would be directly amenable for such big data-driven fields requiring a statistical data analysis. In this paper, we address this problem. We will see that the term ’digital twin’, as used in the literature, is like a Matryoshka doll. For this reason, we unstack the concept via a data-centric machine learning perspective, allowing us to define its main components. As a consequence, we suggest to use the term Digital Twin System instead of digital twin because this highlights its complex interconnected substructure. In addition, we address ethical concerns that result from treatment suggestions for patients based on simulated data and a possible lack of explainability of the underling models. Full article

The latest industrial revolution 4 enabled significant performance improvement through technological advancements. Simultaneously, the industry is setting high-level expectations for changing business practices toward long-term benefits in all three sustainability dimensions. The concept of sustainability embraces all production and operation processes in the Architecture, Engineering, and Construction (AEC) industry. This study systematically explores the literature on sustainability with Enterprise Resource Planning (ERP) and Building Information Modelling (BIM) technologies in the AEC industry and the sustainability vision for their integration. The different types of ERP and BIM implementations have similarities in addressing the broad scope of functionalities. The emergence and proliferation of ERP and BIM have brought crucial changes to the business environment. Further evolution to cloud-based operations is transforming companies from technology-oriented practices to data-centric decision-making smart infrastructures. The narrative literature review investigates the sustainability insights and ideas in ERP and BIM solutions, presenting state of the art on systems integration topics. The relevant literature was retrieved to achieve the research objectives which were qualitatively analyzed to generate the basis for further research. Full article

Electrodialysis (ED) may be designed to enhance wastewater recycling efficiency for crop irrigation in areas where water distribution is otherwise inaccessible. ED process controls are difficult to manage because the ED cells need to be custom-built to meet local requirements, and the wastewater influx often has heterogeneous ionic properties. Besides the underlying complex chemical phenomena, recycling screening is a challenge to engineering because the number of experimental trials must be maintained low in order to be timely and cost-effective. A new data-centric approach is presented that screens three water quality indices against four ED-process-controlling factors for a wastewater recycling application in agricultural development. The implemented unsupervised solver must: (1) be fine-tuned for optimal deployment and (2) screen the ED trials for effect potency. The databionic swarm intelligence classifier is employed to cluster the L₉(3⁴) OA mini-dataset of: (1) the removed Na⁺ content, (2) the sodium adsorption ratio (SAR) and (3) the soluble Na⁺ percentage. From an information viewpoint, the proviso for the factor profiler is that it should be apt to detect strength and curvature effects against not-computable uncertainty. The strength hierarchy was analyzed for the four ED-process-controlling factors: (1) the dilute flow, (2) the cathode flow, (3) the anode flow and (4) the voltage rate. The new approach matches two sequences for similarities, according to: (1) the classified cluster identification string and (2) the pre-defined OA factorial setting string. Internal cluster validity is checked by the Dunn and Davies–Bouldin Indices, after completing a hyper-parameter L₈(4¹2²) OA screening. The three selected hyper-parameters (distance measure, structure type and position type) created negligible variability. The dilute flow was found to regulate the overall ED-based separation performance. The results agree with other recent statistical/algorithmic studies through external validation. In conclusion, statistical/algorithmic freeware (R-packages) may be effective in resolving quality multi-indexed screening tasks of intricate non-linear mini-OA-datasets. Full article

Limited training data is one of the biggest challenges in the industrial application of deep learning. Generating synthetic training images is a promising solution in computer vision; however, minimizing the domain gap between synthetic and real-world images remains a problem. Therefore, based on a real-world application, we explored the generation of images with physics-based rendering for an industrial object detection task. Setting up the render engine’s environment requires a lot of choices and parameters. One fundamental question is whether to apply the concept of domain randomization or use domain knowledge to try and achieve photorealism. To answer this question, we compared different strategies for setting up lighting, background, object texture, additional foreground objects and bounding box computation in a data-centric approach. We compared the resulting average precision from generated images with different levels of realism and variability. In conclusion, we found that domain randomization is a viable strategy for the detection of industrial objects. However, domain knowledge can be used for object-related aspects to improve detection performance. Based on our results, we provide guidelines and an open-source tool for the generation of synthetic images for new industrial applications. Full article

Content-Centric Networking (CCN) is one of the emerging paradigms for the future Internet, which shifts the communication paradigm from host-centric to data-centric. In CCN, contents are delivered by their unique names, and a public-key-based signature is built into data packets to verify the authenticity and integrity of the contents. To date, research has tried to accelerate the validation of the given data packets, but existing techniques were designed to improve the performance of content verification from the requester’s viewpoint. However, we need to efficiently verify the validity of data packets in each forwarding engine, since the transmission of invalid packets influences not only security but also performance, which can lead to a DDoS (Distributed Denial of Service) attack on CCN. For example, an adversary can inject a number of meaningless packets into CCN to consume the forwarding engines’ cache and network bandwidth. In this paper, a novel authentication architecture is introduced, which can support faster forwarding by accelerating the performance of data validation in forwarding engines. Since all forwarding engines verify data packets, our authentication architecture can eliminate invalid packets before they are injected into other CCN nodes. The architecture utilizes public-key based authentication algorithms to support public verifiability and non-repudiation, but a novel technique is proposed in this paper to reduce the overhead from using PKI for verifying public keys used by forwarding engines and end-users in the architecture. The main merit of this work is in improving the performance of data-forwarding in CCN regardless of the underlying public-key validation mechanism, such as PKI, by reducing the number of accesses to the mechanism. Differently from existing approaches that forgive some useful features of the Naive CCN for higher performance, the proposed technique is the only architecture which can support all useful features given by the Naive CCN. Full article