Massive Data Acquisition Research Articles

Designing Heterogeneous LLM Agents for Financial Sentiment Analysis

Large language models (LLMs) have drastically changed the possible ways to design intelligent systems, shifting the focus from massive data acquisition and new model training to human alignment and strategical elicitation of the full potential of existing pre-trained models. This paradigm shift, however, is not fully realized in financial sentiment analysis (FSA), due to the discriminative nature of this task and a lack of prescriptive knowledge of how to leverage existing generative models in such a context. This study investigates the effectiveness of the new paradigm, i.e., using LLMs without fine-tuning for FSA. Rooted in Minsky’s theory of mind and emotions, a design framework with heterogeneous LLM agents is proposed and applied to FSA. The framework instantiates specialized agents using prior guiding knowledge from both linguistics and finance. Then, a summative agent reasons on the aggregated agent discussions. Comprehensive evaluations using six FSA datasets show that the framework yields better accuracies compared to many alternative multi-LLM agent settings, especially when the discussion contents are substantial. This study contributes to the design foundations and paves new avenues for LLMs-based FSA and potentially other tasks. Lastly, implications for business and management have also been discussed.

Portable Arduino-Based Multi-Sensor Device (SBEDAD): Measuring the Built Environment in Street Cycling Spaces.

The built environment's impact on human activities has been a hot issue in urban research. Compared to motorized spaces, the built environment of pedestrian and cycling street spaces dramatically influences people's travel experience and travel mode choice. The streets' built environment data play a vital role in urban design and management. However, the multi-source, heterogeneous, and massive data acquisition methods and tools for the built environment have become obstacles for urban design and management. To better realize the data acquisition and for deeper understanding of the urban built environment, this study develops a new portable, low-cost Arduino-based multi-sensor array integrated into a single portable unit for built environment measurements of street cycling spaces. The system consists of five sensors and an Arduino Mega board, aimed at measuring the characteristics of the street cycling space. It takes air quality, human sensation, road quality, and greenery as the detection objects. An integrated particulate matter laser sensor, a light intensity sensor, a temperature and humidity sensor, noise sensors, and an 8K panoramic camera are used for multi-source data acquisition in the street. The device has a mobile power supply display and a secure digital card to improve its portability. The study took Beijing as a sample case. A total of 127.97 G of video data and 4794 Kb of txt records were acquired in 36 working hours using the street built environment data acquisition device. The efficiency rose to 8474.21% compared to last year. As an alternative to conventional hardware used for this similar purpose, the device avoids the need to carry multiple types and models of sensing devices, making it possible to target multi-sensor data-based street built environment research. Second, the device's power and storage capabilities make it portable, independent, and scalable, accelerating self-motivated development. Third, it dramatically reduces the cost. The device provides a methodological and technological basis for conceptualizing new research scenarios and potential applications.

Clustering and Geodesic Scaling of Dissimilarities on the Spherical Surface

AbstractSpherical embedding is an important tool in several fields of data analysis, including environmental data, spatial statistics, text mining, gene expression analysis, medical research and, in general, areas in which the geodesic distance is a relevant factor. Many data acquisition technologies are related to massive data acquisition, and these high-dimensional vectors are often normalised and transformed into spherical data. In this representation of data on spherical surfaces, multidimensional scaling plays an important role. Traditionally, the methods of clustering and representation have been combined, since the precision of the representation tends to decrease when a large number of objects are involved, which makes interpretation difficult. In this paper, we present a model that partitions objects into classes while simultaneously representing the cluster centres on a spherical surface based on geodesic distances. The model combines a partition algorithm based on the approximation of dissimilarities to geodesic distances with a representation procedure for geodesic distances. In this process, the dissimilarities are transformed in order to optimise the radius of the sphere. The efficiency of the procedure described is analysed by means of an extensive Monte Carlo experiment, and its usefulness is illustrated for real data sets. Supplementary material to this paper is provided online.

Severe viral lower respiratory tract infections in Brazilian children: Clinical features of a national cohort

BackgroundThe accurate etiological diagnosis of lower respiratory tract infections (LRTI) is essential for their effective clinical management. The extensive use of molecular methods during the COVID-19 pandemic has enabled massive data acquisition on viral lower respiratory tract infections. The current study aims to identify clinical features associated with eight viral agents among children presenting severe LRTI. Methodsretrospective cohort study of data from the Brazilian Influenza Epidemiological Surveillance Information System. Patients under 20 years-old who had severe LRTI with etiological confirmation through RT-PCR between 2020 and 2022 were included. Binary logistic regressions were used to examine associations between pathogens and symptoms. Results60,657 cases were assessed. The main viral agents detected were Sars-CoV-2 (COV2) (41.2%), Respiratory Syncytial Virus (29.1%), Human Rhinovirus (HRV) (12.1%), and Influenza (FLU) (5.5%). A general mortality rate of 4.3% was observed. The multivariate analysis evidenced that COV2 less likely presented with cough (OR: 0.34; 95%CI: 0.32–0.36), respiratory discomfort (Adjusted Odds Ratio (aOR): 0.61; 95%Confidence Interval (CI): 0.59–0.64), and desaturation (aOR: 0.71; 95%CI: 0.69–0.75). RSV strongly associated with cough (aOR: 2.59; 95%CI: 2.45–2.75) and respiratory discomfort (aOR: 1.54; 95%CI: 1.46–1.62), whereas FLU was linked to fever (aOR: 2.27; 95%CI: 2.06–2.50) and sore throat (aOR: 1.48; 95%CI: 1.34–1.64). ConclusionsThe viral agents responsible for severe LRTI have distinct associations with clinical features in children.

Transfer condition monitoring across wind turbines using feature alignment and parameter fine-tuning

Transfer condition monitoring across wind turbines still exists two issues. (1) The model of single turbine trained is directly used for other ones in the wind farm to result in low monitoring accuracy because of individual features of each turbine; (2) if the model is trained separately for each turbine, it is not only time-consuming and inefficient, but also is difficult to transfer across turbines. For solving the two issues, a transfer condition monitoring method across wind turbines using feature alignment and parameter fine-tuning is proposed. First, the representative normal turbine in a wind farm is screened by calculating the correlation between the monitoring parameters of wind turbines; second, the input features are constructed by using the massive normal supervisory control and data acquisition data of the representative turbine. The intelligent condition monitoring model of wind turbines is designed by using convolutional neural networks for spatial features and long-term and short-term memory networks for time features. Finally, feature alignment is performed using the normal data of the representative turbine and a large amount of historical normal data from other turbines to reduce data distribution differences. The aligned data is inputted into pre-training monitoring models for parameter fine-tuning. The personalized features of each turbine are incorporated to achieve high-precision transfer condition monitoring across turbines. The proposed method is verified by the real-world turbine data from a wind farm cooperated with us. The results show that the proposed transfer condition monitoring method using feature alignment and parameter fine-tuning can accurately monitor and identify the health states of wind turbines.

Longwave infrared (6.6-11.4 µm) dual-comb spectroscopy with 240,000 comb-mode-resolved data points at video rate.

Using sub-3-cycle pulses from mode-locked Cr:ZnS lasers at λ ≈ 2.4 µm as a driving source, we performed high-resolution dual-frequency-comb spectroscopy in the longwave infrared (LWIR) range. A duo of highly coherent broadband (6.6-11.4 µm) frequency combs were produced via intrapulse difference frequency generation in zinc germanium phosphide (ZGP) crystals. Fast (up to 0.1 s per spectrum) acquisition of 240,000 comb-mode-resolved data points, spaced by 80 MHz and referenced to a Rb clock, was demonstrated, resulting in metrology grade molecular spectra of N2O (nitrous oxide) and CH3OH (methane). The key to high-speed massive spectral data acquisition was low intensity and phase noise of the LWIR combs and high (7.5%) downconversion efficiency, resulting in a LWIR power of 300 mW for each comb.

Unavoidable renaissance of electron metrology in the age of high numerical aperture extreme ultraviolet lithography

Things are drastically changing in the field of metrology. The main reason for that is related to the daunting specification requirements for metrology imposed by high numerical aperture extreme ultraviolet lithography (high NA EUVL). We observe a variety of generation e-beam tools proliferating in imec unique ecosystem, from in-line transmission electron microscope (TEM) to voltage contrast (VC) overlay tools, from die to database (D2DB) large area scanning electron microscope (SEM) to high-voltage SEM, from artificial intelligence (AI)-based inspection tools to massive data acquisition e-beam system. We are facing a renaissance of e-beam metrology. We are going to describe the challenges as well as the latest evolutionary developments of e-beam metrology in the semiconductor industry.

Methodologies to Determine Geometrical Similarity Patterns as Experimental Models for Shapes in Architectural Heritage

Today, plans to protect historic buildings focus on managing architectural heritage sustainably. Technical teams, such as architects and restorers, use massive data acquisition techniques, so an identification mechanism is required to select geometrical similarity patters to support hypothesis that guarantee historical data. Moreover, computational methods are required to understand the role of organic shapes in historic buildings. This paper first describes an extensive review of the literature and then the algorithms and methods to compare and to detect similar geometrical elements and complex patterns in architecture and archaeology. For this purpose, two key aspects are considered: the metric standpoint and historical-graphical features of the 3D models, i.e., composition, techniques, styles, and historical-graphical documentary sources. Research implies testing several methodological lines to know the similarity degree of complex organic shapes in architectural details through statistical analysis, software to assess point clouds, and complex curve analysis. The results have shown that the three procedures can be compared and that the bases of the pillars of both the Cathedral of Seville and the churches in Carmona, Jerez, and Morón are very similar; however, the base of the pillar of the church in Carmona presents scalability variations.

MsWH: A Multi-Sensory Hardware Platform for Capturing and Analyzing Physiological Emotional Signals.

This paper presents a new physiological signal acquisition multi-sensory platform for emotion detection: Multi-sensor Wearable Headband (MsWH). The system is capable of recording and analyzing five different physiological signals: skin temperature, blood oxygen saturation, heart rate (and its variation), movement/position of the user (more specifically of his/her head) and electrodermal activity/bioimpedance. The measurement system is complemented by a porthole camera positioned in such a way that the viewing area remains constant. Thus, the user’s face will remain centered regardless of its position and movement, increasing the accuracy of facial expression recognition algorithms. This work specifies the technical characteristics of the developed device, paying special attention to both the hardware used (sensors, conditioning, microprocessors, connections) and the software, which is optimized for accurate and massive data acquisition. Although the information can be partially processed inside the device itself, the system is capable of sending information via Wi-Fi, with a very high data transfer rate, in case external processing is required. The most important features of the developed platform have been compared with those of a proven wearable device, namely the Empatica E4 wristband, in those measurements in which this is possible.

Splitting and placement of data-intensive applications with machine learning for power system in cloud computing

Aiming to meet the growing demand for observation and analysis in power systems that based on Internet of Things (IoT), machine learning technology has been adopted to deal with the data-intensive power electronics applications in IoT. By feeding previous power electronic data into the learning model, accurate information is drawn, and the quality of IoT-based power services is improved. Generally, the data-intensive electronic applications with machine learning are split into numerous data/control constrained tasks by workflow technology. The efficient execution of this data-intensive Power Workflow (PW) needs massive computing resources, which are available in the cloud infrastructure. Nevertheless, the execution efficiency of PW decreases due to inappropriate sub-task and data placement. In addition, the power consumption explodes due to massive data acquisition. To address these challenges, a PW placement method named PWP is devised. Specifically, the Non-dominated Sorting Differential Evolution (NSDE) is used to generate placement strategies. The simulation experiments show that PWP achieves the best trade-off among data acquisition time, power consumption, load distribution and privacy preservation, confirming that PWP is effective for the placement problem.

Predicting cell behaviour parameters from glioblastoma on a chip images. A deep learning approach

The broad possibilities offered by microfluidic devices in relation to massive data monitoring and acquisition open the door to the use of deep learning technologies in a very promising field: cell culture monitoring. In this work, we develop a methodology for parameter identification in cell culture from fluorescence images using Convolutional Neural Networks (CNN). We apply this methodology to the in vitro study of glioblastoma (GBM), the most common, aggressive and lethal primary brain tumour. In particular, the aim is to predict the three parameters defining the go or grow GBM behaviour, which is determinant for the tumour prognosis and response to treatment. The data used to train the network are obtained from a mathematical model, previously validated with in vitro experimental results. The resulting CNN provides remarkably accurate predictions (Pearson's ρ > 0.99 for all the parameters). Besides, it proves to be sound, to filter noise and to generalise. After training and validation with synthetic data, we predict the parameters corresponding to a real image of a microfluidic experiment. The obtained results show good performance of the CNN. The proposed technique may set the first steps towards patient-specific tools, able to predict in real-time the tumour evolution for each particular patient, thanks to a combined in vitro-in silico approach.

Bulk scanning method of a heavy metal concentration in tailings of a gold mine using SWIR hyperspectral imaging system

This work introduces a bulk data acquisition method using a hyperspectral imaging system (HIS) to measure chromium (Cr) concentration in the soil samples obtained from tailings of a gold mine considering of the spectral competition between heavy metal elements. The chemical, mineralogical, and spectroscopic analyses in a laboratory environment revealed that heavy metal elements' competitive geochemical behaviors were manifested as spectral competitions between heavy metal elements in tailings. The heavy metal elements can be classified into two groups based on their geochemical behaviors: chromium-nickel (Cr-Ni) and zinc-arsenic-cadmium-lead (Zn-As-Cd-Pb). We found an inverse relationship between the two groups in their spectral absorption depth changing patterns, possibly caused by their competition in bonding with the agent minerals. The prediction model of Cr concentrations in the tailing samples using a short-wave infrared (SWIR) HIS was developed from the sample data analysis. The imaging model of Cr concentration is statistically significant with R2 = 0.7 and NRMSE = 11% to 12%. The future use of HIS for massive data acquisition of heavy metal concentration in a natural environment is made possible with such pilot spectroscopic analyses presented in this work.

Guest Editorial: Edge Intelligence for Beyond 5G Networks

Beyond fifth-generation (B5G) networks, or so-called “6G”, is the next-generation wireless communications systems that will radically change how Society evolves. Edge intelligence is emerging as a new concept and has extremely high potential in addressing the new challenges in B5G networks by providing mobile edge computing and edge caching capabilities together with Artificial Intelligence (AI) to the proximity of end users. In edge intelligence empowered B5G networks, edge resources are managed by AI systems for offering powerful computational processing and massive data acquisition locally at edge networks. AI helps to obtain efficient resource scheduling strategies in a complex environment with heterogeneous resources and a massive number of devices, while meeting the ultra-low latency and ultra-high reliability requirements of novel applications, e.g., self-driving cars, remote operation, intelligent transport systems, Industry 4.0, smart energy, e-health, and AR/VR services. By integrating AI functions into edge networks, radio networks become service-aware and resource-aware to have a full insight into the operating environment and can adapt resource allocation/orchestration in a dynamic manner. Despite the potential of edge intelligence, however, many challenges also need to be addressed in this new paradigm. Until now, limited research efforts have been made on edge intelligence for B5G networks.

A short review: Comparisons of high-throughput phenotyping methods for detecting drought tolerance

Drought is a major threat worldwide for crop production, especially due to the rapid climate changes. Current drought solutions involve improving irrigation system, rainwater harvesting, damming, cloud seeding, and changes of cultivation methods. Despite effective, each solution has economic, environmental, and temporal drawbacks. Among all solutions, the most effective, inexpensive and manageable method is the use of drought-tolerant cultivars via plant breeding. However, conventional plant breeding is a time-consuming and laborious task, especially for phenotypic data acquisition of target traits of numerous progenies. High-throughput phenotyping (HTP) is a recently developed method and has potential to overcome the mentioned issues. HTP offers massive, accurate, rapid, and automatic data acquisition in the breeding procedure and can be a breakthrough for developing drought resistant/tolerant cultivars. This study introduces various methods of HTP to detect drought stress, which can accelerate the breeding processes of drought-tolerant cultivars to provide helpful guidelines for breeders and researchers to choose appropriate methods.

Ultra-deepwater seismic plays offshore Brazil — Future drilling off Santos and Campos Basins

Contemporaneous seismic data acquisition in the Santos and Campos Basins offshore Brazil have focused on image characterization of deepwater and ultra-deepwater reservoirs and their relationship with hydrocarbons originating from synrift source rocks. Our interpretation has mapped the stratigraphy of postsalt turbidite reservoirs, and, on the presalt lithology, it has uncovered the underlying synrift sequences that embrace oil-bearing source rocks and the prolific, recently discovered, microbialite carbonate reservoirs. The new phase in geophysical data acquisition and offshore drilling that started in 1999 bolstered the Brazilian offshore petroleum production to record levels. The new, massive, nonexclusive, speculative 2D and 3D data acquisition surveys conducted offshore the Brazilian coast far exceed the amount of all existing cumulative vintage data. Deepwater drilling programs probed the interpreted new prospects. As whole, the modern geophysics data libraries offshore Brazil brought in the technology era to seismic interpretation, reservoir characterization, and geosteering operations in deepwater development drilling. Still, regional interpretation mapping of the outer shelf, slope, deepwater and ultra-deepwater provinces of the Santos and Campos Basins indicates plenty of prospective future drilling in the salt locked minibasins of the ultra-deepwater provinces. Salt tectonics shapes the architecture of these basins; hence, postsalt deepwater turbidite plays were readily interpreted from seismic amplitudes of the modern data that also allow for resolution images of the synrift source rocks, salt architecture, migration paths through faulting and salt windows, reservoir characterization, and regional seal mapping. The new techniques of prestack depth migration have enabled uncovering the imaging structure of the synrift that led to characterization of the presalt carbonate reservoirs and discovery of giant accumulations. Future offshore exploration will continue aiming at postsalt deepwater and ultra-deepwater minibasins plus presalt plays under the massive salt walls, still an underexplored frontier.

Design of a Volumetric Imaging Sequence Using a Vantage-256 Ultrasound Research Platform Multiplexed With a 1024-Element Fully Sampled Matrix Array.

Ultrasound imaging using a matrix array allows real-time multi-planar volumetric imaging. To enhance image quality, the matrix array should provide fast volumetric ultrasound imaging with spatially consistent focusing in the lateral and elevational directions. However, because of the significantly increased data size, dealing with massive and continuous data acquisition is a significant challenge. We have designed an imaging acquisition sequence that handles volumetric data efficiently using a single 256-channel Verasonics ultrasound research platform multiplexed with a 1024-element matrix array. The developed sequence has been applied for building an ultrasonic pupilometer. Our results demonstrate the capability of the developed approach for structural visualization of an ex vivo porcine eye and the temporal response of the modeled eye pupil with moving iris at the volume rate of 30 Hz. Our study provides a fundamental ground for researchers to establish their own volumetric ultrasound imaging platform and could stimulate the development of new volumetric ultrasound approaches and applications.

A Hierarchical IoT Federation Architecture Based on Local Cloud, Platform and Middleware for Massive Context Data Acquisition in Multiple Sensor Networks

A Hierarchical IoT Federation Architecture Based on Local Cloud, Platform and Middleware for Massive Context Data Acquisition in Multiple Sensor Networks

Risk Analysis of Online Lending Platform Based on Bigdata Technology

In recent years, online lending platforms have emerged a lot of insecurity events. How to judge the security state and health degree of the online lending platforms are becoming the most concerned problems for the investors and supervision departments. In this paper, we’ve adopted the massive data acquisition technology. Spark distributed computing, machine learning and other bigdata technologies to analyze the risks of the online lending platform. Then, a risk early warning model based on factor analysis method was proposed. Practice has proved that the model can help to improve the accuracy of the risk prediction to online lending platforms.

Uplink scheduling of MU-MIMO gateway for massive data acquisition in Internet of things

Due to the uplink-centric characteristic of diverse Internet of things (IoT) applications, the wireless connectivity for the IoT should provide a sufficient data rate for the delivery of a huge amount of data traffic toward the uplink direction; furthermore, it needs to satisfy the low complexity and high energy efficiency for the deployment of resource-constrained IoT devices. To this end, we propose an IoT uplink multiple-input and multiple-output (MIMO) scheduling scheme (IoT-UL-MIMO) for which a multi-user (MU)-MIMO capability is implemented in the IoT-gateway device to greatly extend the uplink bandwidth and support the massive data delivery from a plurality of devices toward the gateway. In the IoT-UL-MIMO, a simple user-selection process that selects a proper set of transmitters by just counting the number of successful control-frame transmissions from the devices is used without a channel-sensing mechanism. The aim of this design principle is a mitigation of the system complexity that reduces the energy consumption, which is a key consideration for resource-constrained IoT devices. The simulation results show significant improvements for the throughput and energy efficiency of IoT-UL-MIMO. A duration-length optimization regarding the uplink scheduling phase is also discussed in this paper.

Closed-loop design evolution of engineering system using condition monitoring through internet of things and cloud computing

Flexibility of a manufacturing system is quite important and advantageous in modern industry, which function in a competitive environment where market diversity and the need for customized product are growing. Key machinery in a manufacturing system should be reliable, flexible, intelligent, less complex, and cost effective. To achieve these goals, the design methodologies for engineering systems should be revisited and improved. In particular, continuous or on-demand design improvements have to be incorporated rapidly and effectively in order to address new design requirements or resolve potential weaknesses of the original design. Design of an engineering system, which is typically a multi-domain system, can become complicated due to its complex structure and possible dynamic coupling between domains. An integrated and concurrent approach should be considered in the design process, in particular in the conceptual and detailed design phases. In the context of multi-domain design, attention has been given recently to such subjects as multi-criteria decision making, multi-domain modeling, evolutionary computing, and genetic programing. More recently, machine condition monitoring has been considered for integration into a scheme of design evolution even though many challenges exist for this to become a reality such as lack of systematic approaches and the existence of technical barriers in massive condition data acquisition, transmission, storage and mining. Recently, the internet of things (IoT) and cloud computing (CC) are being developed quickly and they offer new opportunities for evolutionary design for such tasks as data acquisition, storage and processing. In this paper, a framework for the closed-loop design evolution of engineering systems is proposed in order to achieve continuous design improvement for an engineering system through the use of a machine condition monitoring system assisted by IoT and CC. New design requirements or the detection of design weaknesses of an existing engineering system can be addressed through the proposed framework. A design knowledge base that is constructed by integrating design expertise from domain experts, on-line process information from condition monitoring and other design information from various sources is proposed to realize and supervise the design process so as to achieve increased efficiency, design speed, and effectiveness. The framework developed in this paper is illustrated by using a case study of design evolution of an industrial manufacturing system.