Experimental Data Collection Research Articles

Semantic and pragmatic precision in conversational AI systems.

For a conversational agent, to display intelligent interactive behavior implies the ability to respond to the user's intentions and expectations with correct, consistent and relevant actions with appropriate form and content in a timely fashion. In this paper, we present a data-driven analytical approach to embed intelligence into a conversational AI agent. The method requires a certain amount of (ideally) authentic conversational data, which is transformed in a meaningful way to support intelligent dialog modeling and the design of intelligent conversational agents. These transformations rely on the ISO 24617-2 dialog act annotation standard, and are specified in the Dialogue Act Markup Language (DiAML), extended with plug-ins for articulate representations of domain-specific semantic content and customized communicative functionality. ISO 24617-2 is shown to enable systematic in-depth interaction analysis and to facilitate the collection of conversational data of sufficient quality and quantity of instances of interaction phenomena. The paper provides the theoretical and methodological background of extending the ISO standard and DiAML specifications for use in interaction analysis and conversational AI agent design. The expert-assisted design methodology is introduced, with example applications in the healthcare domain, and is validated in human-agent conversational data collection experiments.

Long-Term Impacts of Temperature Gradients on a Concrete-Encased Steel I-Girder Experiment—Field-Monitored Data

This article delves into an investigation of an experimental steel I-girder section encased in concrete and its instrumentation for the purpose of analyzing temperature distributions in concrete bridges. The study aims to examine the effect of differing air temperatures and thermal loads resulting from solar radiation. Structural health monitoring sensors, including a range of temperature sensors, were connected with the girder to monitor the structural performance using the data acquisition system constantly. The collection of experimental data took place during an exceptionally cold season, and the results of the study revealed both the lateral and vertical distribution of thermal gradients, as well as the fluctuations in these gradients over time. On the premise of the accumulated thermal data, empirical equations were suggested in order to forecast peak lateral as well as vertical temperature gradients, in accordance with the girder’s highest daily and lowest mean temperatures. This temperature variation may lead to longitudinal expansion and contraction in the structure during seasonal variations.

Robot grasping system based on deep learning target detection

The traditional robot grasping system often uses fixed point grasping or demonstrative grasping, but with the increase in the diversity of grasping targets and the randomness of poses, the traditional grasping method is no longer sufficient. A robot grasping method based on deep learning target detection is proposed for a high error rate of target recognition and low success rate of grasping in the robot grasping process. The method investigates the robotic arm hand-eye calibration and the deep learning-based target detection and poses estimation algorithm. The Basler camera is used as the visual perception tool of the robot arm, the AUBO i10 robot arm is used as the main body of the experiment, and the PP-YOLO deep learning algorithm performs target detection and poses estimation on the object. Through the collection of experimental data, several grasping experiments were conducted on the diversity of targets randomly placed in the poses under real scenes. The results showed that the success rate of grasping target detection was 94.93% and the robot grasping success rate was 93.37%.

A granular physics-based model for sediment dispersion

Current models of dispersion in marine sediments include losses due to fluid viscosity and shearing at grain-grain contacts. An additional loss mechanism, which is well established in granular physics but is not included in existing models, is the inelasticity of normal compression at grain-grain contacts. Additionally, force-bearing contact networks, often called “force chains,” involve primarily normal compressive forces. These “force chains” are known to play a dominant role in force transmission in granular materials, including wave propagation. Using theoretical analysis of a 1D model “force chain” as well as DEM simulations in higher spatial dimension, we show that this granular mechanics perspective, where forces are transmitted along lossy force chains, may be able to explain salient features of the acoustic properties of marine sediments. In particular, the low and high-frequency behavior of the attenuation coefficient match a large collection of experimental data. [This was supported by the Office of Naval Research.]

Deep Learning at the Physical Layer for Adaptive Terahertz Communications

Wireless communications in the terahertz (THz) band will become a cornerstone of sixth-generation (6G) networks. The THz channel, however, presents several challenges, such as distance-dependent absorption coefficients that can change the bandwidth significantly in case of mobility. Thus, future THz transmitters will have to switch modulation and bandwidth almost continuously. Moreover, using the same transmission scheme can enable adversaries to leverage smart interfering to inflict more damage with less energy expense. To help enable adaptive and secure THz communications, this article presents the first ever experimental study of modulation and bandwidth classification (MBC) at THz frequencies through deep learning (DL) techniques. We have performed an extensive experimental data collection campaign at 120 GHz with different modulation schemes, signal bandwidth (up to 20 GHz), and different signal-to-noise ratio (SNR) levels. We prove for the first time the feasibility and effectiveness of MBC at THz frequencies, with our DL models reaching accuracy up to 78% and 90% in low- and high-SNR conditions. Furthermore, we investigate the memory and latency constraints that need to be satisfied as a function of the signal bandwidth, and propose a boosting technique to improve the inference quality by trading off latency for accuracy. Finally, we experimentally evaluate the latency of our CNN models through FPGA implementation.

A Review of Swelling Effect on Shale Permeability: Assessments and Perspectives

The significant effect of gas sorption induced swelling on shale permeability has been observed through laboratory measurements and explained through permeability models over the past decades. However, there are lab observations that cannot be explained by these models. This knowledge gap prompts this review. Our goal is to form perspectives on how to resolve this gap through assessing the role of swelling on shale permeability. This goal is achieved through the following three steps: (1) collection of experimental shale permeability data measured under both constant effective stress and constant confining pressure conditions; (2) collection and classification of shale permeability models under the influence of gas sorption induced swelling strains; (3) assessments of co-relations between shale permeability data and permeability models. On the basis of all assessments, we conclude that the discrepancies between model predictions and laboratory measurements depend on the relation between bulk and pore swelling strains, pore size scales, and consistencies of strain treatments in the experiments and models. Models assuming that bulk swelling strain is different from pore swelling strain can better explain lab observations than those assuming that they are the same. When the pore size transitions from the micron-scale to the nano-scale, the effect of swelling on shale permeability gradually diminishes. The inconsistencies between how swelling strains are measured in the lab and treated in the models are common in the literatures and affect the discrepancies between model predictions and lab observations. On the basis of these assessments, we form our perspectives: (1) transformation between bulk and pore swelling strains should be characterized and incorporated both for experiments and in permeability models; (2) shale multiscale pore structural characteristics should be characterized when assessing the swelling effect on shale permeability; (3) the time-dependent nature of swelling strain and permeability evolution should be incorporated into future experiments and permeability models.

Experimental and LBM analysis of medium-Reynolds number fluid flow around NACA0012 airfoil

PurposeThe purpose of this paper is the examination of fluid flow around NACA0012 airfoil, with the aim of the numerical validation between the experimental results in the wind tunnel and the Lattice Boltzmann method (LBM) analysis, for the medium Reynolds number (Re = 191,000). The LBM–large Eddy simulation (LES) method described in this paper opens up opportunities for faster computational fluid dynamics (CFD) analysis, because of the LBM scalability on high performance computing architectures, more specifically general purpose graphics processing units (GPGPUs), pertaining at the same time the high resolution LES approach.Design/methodology/approachProcess starts with data collection in open-circuit wind tunnel experiment. Furthermore, the pressure coefficient, as a comparative variable, has been used with varying angle of attack (2°, 4°, 6° and 8°) for both experiment and LBM analysis. To numerically reproduce the experimental results, the LBM coupled with the LES turbulence model, the generalized wall function (GWF) and the cumulant collision operator with D3Q27 velocity set has been used. Also, a mesh independence study has been provided to ensure result congruence.FindingsThe proposed LBM methodology is capable of highly accurate predictions when compared with experimental data. Besides, the special significance of this work is the possibility of experimental and CFD comparison for the same domain dimensions.Originality/valueConsidering the quality of results, root-mean-square error (RMSE) shows good correlations both for airfoil’s upper and lower surface. More precisely, maximal RMSE for the upper surface is 0.105, whereas 0.089 for the lower surface, regarding all angles of attack.

Effect of Stress Interaction on Multi-Stress Accelerated Life Test Plan: Assessment Based on Particle Swarm Optimization

Sustainability of products that seek to maintain ecosystem balance, such as electric vehicles or solar system inverters, often require extensive testing during their developmental stages in a manner that minimizes wastage and drives creativity. Multi-stress accelerated test planning is often used for these products, their subsystems and components if their in-service failures are driven by multiple stress factors. Multi-stress accelerated life testing (ALT) often expedites time to failure for highly reliable products. Many studies assume model parameters that may not be appropriate for the considered stress factors. Most importantly, the effect stress interaction has on the ALT plan is often ignored, especially for cases where historical data are lacking. To address this gap, in this work, a technique based on a combination of rapid experimental data collection and heuristic-based optimization is proposed for ALT planning. In addition, the effect of stress interaction on the ALT plan was also evaluated. Specifically, the Arrhenius model was used to develop a maximum likelihood mathematical expression for multi-stress factor scenarios with and without interaction. Subsequently, two optimization stages based on particle swarm optimization (PSO) were carried out using time varying inertia weight constants to drive early and late global and local searches, respectively. In the first stage, model parameters were estimated, while, in the second stage, an ALT optimal plan was generated based on a D-optimality criterion. Verification of stress factor interactions was carried out using graphical response analysis. An experiment, designed to investigate electromigration in solder joints under three stress factors (temperature, current density and mechanical load), was used to validate the study. The variation in the choice of Latin hypercube design (LHD) results in disparity in the levels of stress within each stress combination as well as sample allocation. Our results clearly show the need to investigate stress interactions prior to multi-stress acceleration planning.

Charged particle reconstruction in CLAS12 using Machine Learning

In this work, we present studies of track parameter reconstruction from raw information in CLAS12 detector's Drift Chambers, using Machine Learning (ML). We study the resolution of tracks reconstructed with different types of ML models/algorithms, including Multi-Layer Perceptron (MLP), Extremely Randomized Trees (ERT) and Gradient Boosting Trees (GBT) using simulated data. The resulting ML model is capable of reconstructing track parameters (particle momentum, and polar and azimuthal angles) with accuracy similar to Hit Based (HB) tracking code, but 150 times faster. Moreover, physics reactions can be identified using the particles reconstructed by the neural network in real-time (with a rate of about 34kHz) during experimental data collection. The developed model can be used in numerous applications, such as triggering specific physics reactions in real-time, detector performance monitoring, and real-time detector calibration.

Decision Strategy Tool for the Design of Natural Treatment Systems for Wastewater (NTSW) from Isolated Livestock Farms

This work proposes a series of strategies and tools for the design and characterization of natural treatment systems of wastewater (NTSW) applied in livestock farms, considering the parameters of flow rate (Q), conductivity (EC) and chemical oxygen demand (COD) of the waste generated, farm location, hydraulic retention time (HRT), and removal/reduction rate targets. The high organic load of the effluent generated in these farms has an important environmental impact, which is amplified in insular or isolated territories. The application of such treatment systems has demonstrated their suitability in these environments, but their design lacks proper characterization and sizing tools for their adequate operation. The proposed tools in this work are based on a collection of experimental data over a ten-year period of application of NTSW in real farms. This work contributes to facilitate the design and implementation of NTSW in farms located in isolated, island, or similar-size environments. Finally, as a practical application, an inventory and implementation of the tool developed for livestock farms on the island of Gran Canaria (Spain) is carried out.

Configuration Information System for online processing and data monitoring in the NICA experiments

The collection, storage and processing of experimental data are an integral part of modern high-energy physics experiments. The Configuration Information System (CIS) based on the Configuration Database (ConfigDb) is an essential element of a complex of information systems developed for the NICA experiments. The CIS is used to store and provide data on the configuration of the experiment hardware and software systems when collecting data from the detectors in online mode. The ConfigDb stores both a set of various configuration parameters and descriptions of a sequence of online software tasks to be started and run during experiment sessions. The system allows loading configuration information, activating hardware setups and launching all necessary software applications with the required parameters on the specified distributed nodes. The CIS architecture mainly contains the ConfigDb and Configuration Manager, which uses the Dynamic Deployment System API for managing tasks and providing intercommunication. The manager loads configuration data from the database, starts software tasks in a described sequence and supports task process management. The client-server architecture of the CIS is presented. The client has been implemented as a Web service to manage configuration parameters by users.

GROWTH RESPONSE OF PAKCOY (Brassica rapa L.) WITH HYDROPONIC APPLICATION OF LIQUID ORGANIC FERTILIZER FRUIT WASTE

Research on the application of fruit waste liquid organic fertilizer to the growth of pakcoy plants (Brassica rapa L.) has been carried out. The purpose of this study was to determine the growth response of the pakcoy plant (Brassica rapa L.) to the application of liquid organic fertilizer from fruit waste at various concentrations hydroponically and to determine the best concentration of liquid organic fertilizer for the research results. The design of this study used CRD (Completely Randomized Design) with experimental data collection methods. This research was conducted with one control treatment, 5 treatments and 4 replications namely D0 (100% AB mix), D1 (25%), D2 (40%), D3 (55%), D4 (70%) and D5 (85%). Data were analyzed statistically through ANOVA test and followed by 5% DMRT test. Based on the research results obtained, the application of liquid organic fertilizer from fruit waste had a significant effect on the growth response of pakcoy plants (Brassica rapa L.). This can be seen in the best treatment found in treatment D5 (85%), as evidenced by the highest average result of fruit waste liquid organic fertilizer at plant height: 22.075 cm, number of leaves: 20.25 leaves and fresh weight: 41. 25 gr/plant. Thus, the use of fruit waste as a POC affects the effectiveness of pakcoy growth.

Imaging optimisation process of human lung tissue using synchrotron radiation microcomputed tomography in PLS‐II

AbstractSynchrotron radiation microcomputed tomography (SR‐μCT) can provide non‐destructive three‐dimensional investigation for soft tissue with high resolution. The aim of this study is to evaluate technical improvements in microcomputed tomographic imaging of human lung tissue using a synchrotron radiation source at the 6C biomedical beamline. SR‐μCT images of excised human lung tissue (6 mm vertically and 10 mm horizontally sized) were explored at the 6C biomedical beamline in the Pohang accelerator laboratory (Pohang, Republic of Korea) using phase retrieval methods with a modified Bronnikov algorithm. Experimental data collection was performed with changes in the following conditions: the presence of a graphite attenuator, sample to detector distance, the presence of different scintillators and increasing pixel sizes. We identified that the resolution of lung tissue images was sufficiently enhanced when the graphite attenuator was not applied, 550 mm of sample to detector distance, with Cadmium tungsten (CdWO4) scintillator and twice binning of the pixel data and could be comparable with images from pathologic examination. Refinement of SR‐μCT techniques for soft tissue originating from the human body would advance the potential clinical application of synchrotron radiation in the near future.

Federated Meta-Learning with Attention for Diversity-Aware Human Activity Recognition.

The ubiquity of smartphones equipped with multiple sensors has provided the possibility of automatically recognizing of human activity, which can benefit intelligent applications such as smart homes, health monitoring, and aging care. However, there are two major barriers to deploying an activity recognition model in real-world scenarios. Firstly, deep learning models for activity recognition use a large amount of sensor data, which are privacy-sensitive and hence cannot be shared or uploaded to a centralized server. Secondly, divergence in the distribution of sensory data exists among multiple individuals due to their diverse behavioral patterns and lifestyles, which contributes to difficulty in recognizing activity for large-scale users or 'cold-starts' for new users. To address these problems, we propose DivAR, a diversity-aware activity recognition framework based on a federated Meta-Learning architecture, which can extract general sensory features shared among individuals by a centralized embedding network and individual-specific features by attention module in each decentralized network. Specifically, we first classify individuals into multiple clusters according to their behavioral patterns and social factors. We then apply meta-learning in the architecture of federated learning, where a centralized meta-model learns common feature representations that can be transferred across all clusters of individuals, and multiple decentralized cluster-specific models are utilized to learn cluster-specific features. For each cluster-specific model, a CNN-based attention module learns cluster-specific features from the global model. In this way, by training with sensory data locally, privacy-sensitive information existing in sensory data can be preserved. To evaluate the model, we conduct two data collection experiments by collecting sensor readings from naturally used smartphones annotated with activity information in the real-life environment and constructing two multi-individual heterogeneous datasets. In addition, social characteristics including personality, mental health state, and behavior patterns are surveyed using questionnaires. Finally, extensive empirical results demonstrate that the proposed diversity-aware activity recognition model has a relatively better generalization ability and achieves competitive performance on multi-individual activity recognition tasks.

Data collection experiments with multiple robots considering function and location

In this study, we proposed the networked system for collaboration with human and robots, considering location and robot function. We placed several robots in the museum and conducted data collection experiments for a coexisting society with human and robots to cope with the shortage of human resources and social issues. As a communication system, we built a platform considering the location and function of the robots. 7 robots and 6 tablets were connected to the platform and networked. The data were stored in a database and could be checked the robot and event situations from the web page. In the future, we will continue to build the database and server process for service in consideration of connecting more types of robots and sensors to the platform.

Development of an Animal Farm Robotic Feeding System

Feeding of farm animals is an essential factor for animal growth and productivity. The traditional manual feeding systems of farm animals have been posed with lots of challenges ranging from the high cost of manpower to poor feed management. Existing automatic feeding systems are expensive and are usually fixed systems. This paper presents an autonomous robotic system that collects and delivers feed to farm animals. The system design consists of an automatic feed dispensing system and an unmanned ground robotic vehicle. The vehicle collects feeds from the feed dispensing unit for delivery at the animal pen house in a soft real-time order. An enclosed environment, as is found in many farm animal production settings, is used to enable auto-navigation of the robot system within the farm. Auto-navigation, feed collection and delivery functions are implemented using ultrasonic sensor and infrared (IR) proximity sensors in a microcontroller (Atmega328p) based hardware and firmware control algorithm. Control parameter variation was achieved through Bluetooth (HC-05) interface communication with a mobile device. Experimental test and data collection from the system indicates the system can autonomously supply 26.6kg worth of feed within 1hr at a speed of 1.233km/h and apply to feeding various kinds of farm animals.

Deep learning reveals what vocal bursts express in different cultures.

Human social life is rich with sighs, chuckles, shrieks and other emotional vocalizations, called 'vocal bursts'. Nevertheless, the meaning of vocal bursts across cultures is only beginning to be understood. Here, we combined large-scale experimental data collection with deep learning to reveal the shared and culture-specific meanings of vocal bursts. A total of n = 4,031 participants in China, India, South Africa, the USA and Venezuela mimicked vocal bursts drawn from 2,756 seed recordings. Participants also judged the emotional meaning of each vocal burst. A deep neural network tasked with predicting the culture-specific meanings people attributed to vocal bursts while disregarding context and speaker identity discovered 24 acoustic dimensions, or kinds, of vocal expression with distinct emotion-related meanings. The meanings attributed to these complex vocal modulations were 79% preserved across the five countries and three languages. These results reveal the underlying dimensions of human emotional vocalization in remarkable detail.

Experimental Data Collection of Surface Quality Analysis of CuCrZr Specimens Manufactured with SLM Technology: Analysis of the Effects of Process Parameters

Selective laser melting (SLM) is the most widely used laser powder-bed fusion (L-PBF) technology for the additive manufacturing (AM) of parts from metallic powders. The surface quality of the SLM parts is highly dependent on many factors and process parameters. These factors include the powder grain size, the layer thickness, and the building angle. This paper conducted an experimental analysis of the effects of SLM process parameters on the surface quality of CuCrZr cubic specimens. Thanks to its excellent thermal and mechanical properties, CrCrZr has become one of the most widely used materials in SLM technology. The specimens have been produced with different combinations of layer thickness, laser patterns, building angles, and scanning speed, keeping the energy density constant. The results show how different combinations of parameters affect the surface quality macroscopically (i.e., layer thickness, building angle, and scanning speed); in contrast, other parameters (i.e., laser pattern) do not seem to have any contributions. By varying these parameters within typical ranges of the AM machine used, variations in surface quality can be achieved from 10.4 µm up to 40.8 µm. These results represent an important basis for developing research activities that will further focus on implementing a mathematical/experimental model to help designers optimize the surface quality during the AM pre-processing phase.

Combining diagnostics, modeling, and control systems for automated alignment of the TES beamline

X-ray beamlines are essential components of all synchrotron light sources. Practical operations involve frequent variation in beamline component positions and orientation, particularly when photon beam parameters shift due to experimental needs, or due to variations in the incoming photon beam. The alignment process can be time consuming and takes away from valuable beam time for experimental data collection. We describe progress in the automation of certain alignment tasks on the tender-energy X-ray spectroscopy (TES) beamline at the National Synchrotron Light Source II (NSLS-II). The beamline is controlled using the BlueSky software in which high level experimental plans guide the beamline components during an experiment. Numerous software packages exist for beamline modeling, and they may be tied to the beamline control system using a package we are continuing to develop called Sirepo-Bluesky. The photon beam distribution may be measured with fluorescent screens, and a relation between beam and machine state can be found by varying the mirror and aperture settings over a multi-dimensional range. We describe the results of such parameter varying measurements and how we are combining Sirepo-Bluesky with machine learning methods and reduced models to automate mirror alignment on the TES beamline.

Construction and application of time-effect assessment database for experiments on war-traumatized animals

Abstract Background High-quality real war trauma data are important, but they are still lacking. Using the database to realize the networked, systematic, standardized, and dynamic management of war injury animal experimental data can provide a real and reliable supplement for the research of war injury treatment. Methods Through demand analysis, module design, architecture design, compilation and debugging, and other processes, JAVA language development tools were used to develop a time-effect assessment database for experiments on war-traumatized animals, where the browser/server architecture (B/S architecture) is used as the technical architecture and SQL Server as the database management system. Results A time-effect assessment database for experiments on war-traumatized animals that features reasonable structural design, rich data resources, and convenient use has been established, in which case efficient collection, storage, display, retrieval, and management of animal experimental data have been achieved. At present, the database is running well and has good data quality.