Error Rate Measurements Research Articles

An objective evaluation of the effects of recording conditions and speaker characteristics in multi-speaker deep neural speech synthesis

Multi-speaker spoken datasets enable the creation of text-to-speech synthesis (TTS) systems which can output several voice identities. The multi-speaker (MSPK) scenario also enables the use of fewer training samples per speaker. However, in the resulting acoustic model, not all speakers exhibit the same synthetic quality, and some of the voice identities cannot be used at all.In this paper we evaluate the influence of the recording conditions, speaker gender, and speaker particularities over the quality of the synthesised output of a deep neural TTS architecture, namely Tacotron2. The evaluation is possible due to the use of a large Romanian parallel spoken corpus containing over 81 hours of data. Within this setup, we also evaluate the influence of different types of text representations: orthographic, phonetic, and phonetic extended with syllable boundaries and lexical stress markings.We evaluate the results of the MSPK system using the objective measures of equal error rate (EER) and word error rate (WER), and also look into the distances between natural and synthesised t-SNE projections of the embeddings computed by an accurate speaker verification network. The results show that there is indeed a large correlation between the recording conditions and the speaker’s synthetic voice quality. The speaker gender does not influence the output, and that extending the input text representation with syllable boundaries and lexical stress information does not equally enhance the generated audio across all speaker identities. The visualisation of the t-SNE projections of the natural and synthesised speaker embeddings show that the acoustic model shifts some of the speakers’ neural representation, but not all of them. As a result, these speakers have lower performances of the output speech.

基于模糊PID路径跟踪算法的农业温室作业机器人智能控制技术

With the development of agricultural automation, applying intelligent algorithms to the navigation control of agricultural work vehicles has important practical significance for improving vehicle navigation accuracy and operation efficiency. In view of the complexity of the agricultural greenhouse environment, this study proposed a fuzzy PID path tracking algorithm based on the traditional vehicle PID control system. This algorithm uses a fuzzy controller to improve the PID control system, thereby realizing the online setting of PID control parameters. In order to verify the effectiveness of the fuzzy PID path tracking algorithm, the improved control system was applied to the tracked vehicle robot of Beijing Forestry University, and the operation performance of the vehicle robot was tested. The research results show that the absolute error rate of vehicle robot distance measurement is less than 1%; the error of the man-machine follow-up test is between 4 and 7 cm, and the measured follow-up distance is slightly less than the safe follow-up distance; the maximum error of the vehicle's fixed-point parking is 0.3 cm; The linear position tracking control has a lateral position deviation of ±3cm, and the vehicle's linear driving control and steering effects are better. The fuzzy PID path tracking algorithm designed this time shows good control performance, which has reference significance for the practical application of agricultural robots.

Influence of flow disturbances on the performance of industry-standard LNG flow meters

In the last decade significant progress has been achieved in the development of measurement traceability for LNG inline metering technologies such as Coriolis and ultrasonic flow meters. In 2019, the world's first LNG research and calibration facility has been realised thus enabling calibration and performance testing of small and mid-scale LNG flow meters under realistic cryogenic conditions at a maximum flow rate of 200 m3/h and provisional mass flow measurement uncertainty of 0.30% (k = 2) using liquid nitrogen as the calibration fluid. This facility enabled, for the first time, an extensive test programme of LNG flow meters under cryogenic conditions to be carried out to achieve three main objectives; the first is to reduce the onsite flow measurement uncertainty for small and mid-scale LNG applications to meet a target measurement uncertainty of 0.50% (k = 2), the second is to systematically assess the impact of upstream flow disturbances and meter insulation on meter performance and the third is to assess transferability of meter calibrations with water at ambient conditions to cryogenic conditions. SI-traceable flow calibration results from testing six LNG flow meters (four Coriolis and two ultrasonic, see acknowledgment section) with water in a water calibration facility and liquid nitrogen (LIN) in the LNG research and calibration facility under various test conditions are fully described in this paper. Water and LIN calibration data were compared and it was observed that the influence of removing the meter insulation on mass flow rate measurement accuracy can be more significant (meter error > ±0.50%) than the influence of many typical upstream disturbances when the meter is preceded by a straight piping length equal to twenty pipe diameters (20D) with no additional flow conditioning devices, in particular for ultrasonic meters. The results indicate that the correction models used to transfer the water calibration to cryogenic conditions (using LIN) can potentially result in mass flow rate measurement errors below ±0.5%, however, the correction models are specific to the meter type and manufacturer. This work shows that the target measurement uncertainty of 0.50% can be achieved if the expanded standard error of the mean value measured by the meter is smaller than 0.40% (k = 2). It is planned to repeat these tests with LNG in order to compare the results with the LIN tests presented in this paper. This may reveal that testing with an explosion safe and environmentally friendly fluid such as LIN produces representative results for testing LNG flow meters.

Prediction of solar particle events with SRAM-based soft error rate monitor and supervised machine learning

This work introduces an embedded approach for the prediction of Solar Particle Events (SPEs) in space applications by combining the real-time Soft Error Rate (SER) measurement with SRAM-based detector and the offline trained machine learning model. The proposed approach is intended for the self-adaptive fault-tolerant multiprocessing systems employed in space applications. With respect to the state-of-the-art, our solution allows for predicting the SER 1 h in advance and fine-grained hourly tracking of SER variations during SPEs as well as under normal conditions. Therefore, the target system can activate the appropriate mechanisms for radiation hardening before the onset of high radiation levels. Based on the comparison of five different machine learning algorithms trained with the public space flux database, the preliminary results indicate that the best prediction accuracy is achieved with the recurrent neural network (RNN) with long short-term memory (LSTM).

MDER: Multi-Dimensional Event Recommendation in Social Media Context

AbstractEvents represent a tipping point that affects users’ opinions and vary depending upon their popularity from local to international. Indeed, social media offer users platforms to express their opinions and commitments to events that attract them. However, owing to the volume of data, users are encountering a difficulty to accede to the preferred events according to their features that are stored in their social network profiles. To surmount this limitation, multiple event recommendation systems appeared. Nevertheless, these systems use a limited number of event dimensions and user’s features. Besides, they consider users’ features stored in a single user’s profile and disregard the semantic concept. In this research, an approach for multi-dimensional event recommendation is set forward to recommend events to users resting on several event dimensions (engagement, location, topic, time and popularity) and some user’s features (demographic data, position and user’s/friend’s interests) stored in multi-user’s profiles by considering the semantic relationships between user’s features, specifically user’s interests. The performance of our approach was assessed using error rate measurements (mean absolute error, root mean squared error and cross-validation). Experiment that results on real-world event data sets confirmed that our approach recommends events that fit the user more than the previous approaches with the lowest error rate values.

Body-Posture Recognition by Undergraduate Students Majoring in Physical Education and Other Disciplines.

Humans are more proficient at processing visual display of body posture when the body is in upright orientation, compared to when inverted (inversion effect). Here we investigated whether extensive exposure or expertise on body posture recognition would affect the efficiency with which body-posture is processed. Using whole-body and piecemeal-body postures as stimuli, we performed two experiments to investigate whether body-posture recognition differed between two groups of participants: undergraduates majoring in physical education (PE) and those in other subjects (non-PE), respectively. These two groups differed significantly in the frequency and intensity of exercise per day and/or accumulated exercise time. In our experiments, following initial presentation of an image of a body posture, participants were shown the same or a different stimulus and were asked to report whether or not they had been previously shown the same image. The orientations of the body postures were also varied between trials. Our results showed that, in Experiment 1, for whole-body posture recognition, both the PE and non-PE groups showed a robust body-inversion effect in terms of both error rate and reaction time (RT), but the magnitude of the body-inversion effect in the RT measure was greater in the PE than the non-PE group. In Experiment 2, for piecemeal-body postures, both groups showed the inversion effect in terms of both error rate and RT measures and the PE group made fewer overall errors than the non-PE group. These cumulative results suggest that a superiority effect exists for PE participants compared with non-PE participants. Our results are generally consistent with the expertise hypothesis.

The performance of permutations and exponential random graph models when analyzing animal networks

Abstract Social network analysis is a suite of approaches for exploring relational data. Two approaches commonly used to analyze animal social network data are permutation-based tests of significance and exponential random graph models. However, the performance of these approaches when analyzing different types of network data has not been simultaneously evaluated. Here we test both approaches to determine their performance when analyzing a range of biologically realistic simulated animal social networks. We examined the false positive and false negative error rate of an effect of a two-level explanatory variable (e.g., sex) on the number and combined strength of an individual’s network connections. We measured error rates for two types of simulated data collection methods in a range of network structures, and with/without a confounding effect and missing observations. Both methods performed consistently well in networks of dyadic interactions, and worse on networks constructed using observations of individuals in groups. Exponential random graph models had a marginally lower rate of false positives than permutations in most cases. Phenotypic assortativity had a large influence on the false positive rate, and a smaller effect on the false negative rate for both methods in all network types. Aspects of within- and between-group network structure influenced error rates, but not to the same extent. In "grouping event-based" networks, increased sampling effort marginally decreased rates of false negatives, but increased rates of false positives for both analysis methods. These results provide guidelines for biologists analyzing and interpreting their own network data using these methods.

DESIGN OF MECHANISM AND MOTION SYSTEM ON TANK PROTOTYPE USING ODOMETRY

At present, Indonesia's defense forces are not yet on par with European countries. The size is the variation of Indonesia's defense equipment technology that has not been able to keep up with the defense equipment of European countries . Utilization of unmanned tank rides is a solution in empowering mobile combat equipment as a means of national defense and defense . Unarmed armored tank defense equipment as a means of national defense and defense . The Odometry technique applied in this study uses an encoder sensor device mounted on the two right and left wheels of the prototype tank. Wheel movement will produce the number of turns that can be calculated using an encoder sensor . the authors use odometry techniques aimed at estimating distances to destinations within a relatively small measurement area, in the wider measuring area this study uses GPS. Another goal is to get the accuracy and accuracy of distance measurements against the destination location . The results obtained in this study are the level of accuracy of the measurement of distance traveled in units of cm from the results of testing and implementation of the odometry system . The application of odometry is used to get the calculation of the change in distance due to the displacement of the prototype location during the autopilot motion process in relation to getting the accuracy of moving the position towards the waypoint . The measurement error rate obtained is less than 1%. It is expected that this system can be used as a motion mechanism system on alutsi s ta tank.
 
 Keywords : Odometri, Encoder , Waypoint, Autopilot

Testing and improving type 1 error performance of Şen’s innovative trend analysis method

Recent Şen innovative trend analysis method (Şen_ITA) has high graphical visual ability in addition to quantitative trend identification analysis aspects. However, it is important to determine critical trend values as to whether they are significant according to a certain statistical significance level. In order to achieve such a goal, a certain amount or average percentages (5%, 10%, and 20%), bootstraps, and variance correction methods are frequently used by the researchers as evident from the literature. These methods accept either approximate critical values or require the transformation of the time series to suit their assumption requirements. In this paper, Monte Carlo simulation studies are used to measure error rates (type 1) as suggestion for critical trend value determinations for the Şen_ITA method. Şen_ITA critical trend formula has been developed originally depending on type 1 error rate exceedance expectation. The improved method (ITA_R) as suggested in this paper provides successful results by comparing it with the classic Mann-Kendall (MK) method. Furthermore, the ITA_R and MK methods have been applied to the mean daily maximum temperature series from England. The resulting trend values are generally consistent with the MK method and show an upward trend in all regions of England and during all seasons.

Characteristics of Reported Pediatric Medication Errors in Northern Ireland and Use in Quality Improvement.

To protect children from harm, clinicians, educators, and patient safety champions need information to direct improvement efforts. Critical incident data could provide this but are often disregarded as a source of evidence because under-reporting makes them an inaccurate measure of error rates. Our aim was to identify key targets for pediatric healthcare quality improvement. The objective was to evaluate the types, characteristics, and areas of risk within reported medication errors in pediatric patients. We conducted a retrospective study of a large regional dataset of 1522 pediatric medication errors reported from secondary care between 2011 and 2015, including all hospitals and community pediatric settings in Northern Ireland. The following characteristics were included: error severity, patient age, drug involved, error type, and area of practice. Two academic pediatricians, a senior medicines governance pharmacist, a Reader in Pharmacy Practice, and a Professor of Medical Education analyzed the data. Validity checks included comparing the findings against key published literature and discussion by a practitioner panel representing five multidisciplinary stakeholder groups. Neonates, particularly in intensive care, were implicated in 19% of all errors. The medications most represented in risk were antimicrobials, paracetamol, vaccines, and intravenous fluids. The error types most implicated were dosing errors (32%) and omissions (21%). Incident reports identified neonates, a shortlist of drugs, and specific error types, associated with modifiable behaviors, as priority improvement targets. These findings direct further study and inform intervention development, such as specific training in calculations to prevent dosing errors. Involving experienced practitioners both endorsed the findings and engaged the practice community in their future implementation. The utility of incident reports to direct improvement efforts may offset the limitations in their representativeness.

Utilizing single-wavelength for OFDM wireless downstream and remodulated OOK upstream in colorless access network to mitigate Rayleigh backscattering noise

Abstract To avoid the Rayleigh backscattering (RB) beat noise in colorless wavelength division multiplexing (WDM) access architecture, the dual-band modulation diagram is proposed for demonstration. In the study, 11.25 Gbit/s orthogonal frequency division multiplexing (OFDM) wireless downstream and 2.5 Gbit/s remodulated on–off keying (OOK) upstream traffic are achieved by using same wavelength. Moreover, the bit error rate (BER) measurements of wireless downstream and wired upstream wavelengths are also executed and discussed after 20 km fiber connection.

Least Loss: A simplified filter method for feature selection

Identifying the relevant set of features in a dataset is an important part of data analytics. Discarding significant variables or keeping irrelevant variables has significant effects on the performance of the learning algorithm during knowledge discovery. In this paper, a feature selection method called Least Loss (L2) is proposed that significantly reduces the dimensionality of data by disposing weakly correlated variables in a robust manner without diminishing the predictive performance of classifiers. The proposed method is based on quantifying the similarity between the observed and expected probabilities and generating scores for each independent variable, which makes it simple and intuitive. The evaluation of the proposed method was done by comparing its performance against Information Gain (IG) and Chi Square (CHI) feature selection methods on 27 different datasets modeled using a probabilistic classifier. The results reveal that L2 is highly competitive with respect to error rate, precision, and recall measures while substantially reducing the number of selected variables in the datasets. Our study would be of high interest to data analysts, scholars and domain experts who deal with applications that include large numbers of features using statistical analysis methods.

Performance Analysis of ZCC-Optical-CDMA over SMF for Fiber-To-The-Home Access Network

To provide a sustainable fiber-to-the-home (FTTH), several multiplexing techniques have been developed for this purpose. The correlation features are the main obstacle behind the network performance limitation, which imposes to a high level of multiple access interference. However, the development of multiplexing techniques helps to overcome these limitations, such as optical-code division multiple access (Optical-CDMA). Optical-CDMA is considered as one of the most powerful solutions for FTTH. This paper aims to enhance FTTH network performance by applying Zero cross-correlation code (ZCC) with Optical-CDMA with maximum link single-mode fiber. In the simulation, the system performance is demonstrated in terms of bit error rate, Q-Factor and eye diagram measurements.

Performance of Low-Noise Ferrite Shield in a K-Rb-21Ne Co-Magnetometer

Spin-exchange relaxation-free (SERF) co-magnetometer is considered to be a promising nuclear spin gyroscope due to its ultrahigh rotation measurement sensitivity and magnetic field disturbance cancellation ability. Recent experiments indicate that the relaxation of the noble-gas spins would result in reduction of the magnetic field disturbance cancellation ability, which means that magnetic noise from the passive magnetic shield in a SERF co-magnetometer can cause rotation rate measurement error. In this paper, the magnetic field frequency response of a SERF atomic magnetometer is derived based on state space method and the magnetic noise of different shield materials are measured. The magnetic noise error model in a SERF co-magnetometer is then derived and the magnetic noise induced rotation rate measurement error is quantitatively evaluated. The superiority of ferrite shield is shown by the performance of a K-Rb-21Ne co-magnetometer. Experimental results show that ferrite shield with lower magnetic noise and better temperature stability is a more favorable choice for the inner shield of the co-magnetometer.

Detecting WDM visible light signals by a single multi-color photodiode with MIMO processing

For the first time, to the best of our knowledge, we experimentally demonstrate that multiple-input-multiple-output (MIMO) processing allows using a single photodiode to detect simultaneously a wavelength-division multiplexing (WDM) visible light communications (VLC) signal. The photodiode has a triple junction, and when it is illuminated by a WDM signal, the junctions produce inherently three photocurrents that are unusable for detecting any of the WDM signals. However, by means of linear MIMO processing, we are able to recover the transmitted signals exactly. Bit error rate measurements confirm the effectiveness of the proposed solution. This opens a new scenario for practical WDM-VLC systems.

Automatic Semantic Segmentation with DeepLab Dilated Learning Network for Change Detection in Remote Sensing Images

Automatic change detection is an interesting research area in remote sensing (RS) technology aims to detect the changes in synthetic aperture radar (SAR) and multi-temporal hyperspectral images acquired at different time intervals. This method identifies the differences between the images and accomplishes the classification result into changed and unchanged areas. However, the existing algorithms are degraded due to noises present in the RS images. The main aim of the proposed method is the automatic semantic segmentation based change detection that produces a final change between the two input images. This paper proposes a feature learning method named deep lab dilated convolutional neural network (DL-DCNN) for the detection of changes from the images. The proposed approach consists of three stages: (i) pre-processing, (ii) semantic segmentation based change detection and (iii) accuracy assessment. Initially, preprocessing is performed to correct the errors and to obtain detailed information from the scene. Then, map the changes between the two images with the help of a trained network. The DCNN network performs fine-tuning and determines the relationship between two images as changed and unchanged pixel areas. The experimental analysis conducted on various datasets and compared with several existing algorithms. The experimental analysis is performed in terms of F-score, percentage correct classification, kappa coefficient, and overall error rate measures to show a better performance measure than the other state-of-art approaches.

Sparse Index OFDM Modulation for IoT Communications

One of the greatest challenges facing the physical layer design of the internet of things (IoT) resides in the imposed constraint of very low power consumption. Recently, new modulation scheme termed OFDM with sparse index modulation (OFDM-SIM) has been introduced as an energy efficient multicarrier scheme (MCS). Although of its high energy efficiency (EE) and spectral efficiency (SE), OFDM-SIM cannot fulfill the IoT energy requirements owing to its high PAPR. In this regard, an enhanced OFDM-SIM is proposed in this paper as an energy efficient MCS for IoT communications. In particular, a novel clipping-compressive sensing (CS) based PAPR reduction technique for OFDM-SIM is proposed. In the transmitter (TX) side, considering the complexity constraints for IoT devices, the simple and low complex clipping method is exploited to deal with the PAPR issue. On the receiver (RX) side, a robust CS signal recovery scheme is proposed to deal with tough resulting clipping noise. Unlike high complex conventional CS-based schemes, the proposed scheme exploits the inherent sparsity of the received enhanced OFDM-SIM signal rather than clipping noise sparsity to achieve a low complex CS signal detection. Moreover, in this paper, the information-theoretic limits on sparsity recovery are exploited to derive an upper bound measure of the bit error rate (BER). The simulation results demonstrate the superiority of the proposed scheme, as it significantly enhances the overall system performance in terms of EE and PAPR reduction compared to the conventional clipped coded-OFDM.

Use of firefly optimization algorithm for fractional fourier transform based reversible data hiding

This paper presents a Fractional Fourier transform based reversible data hiding technique to secure the data transmitted over communication network. The proposed algorithm modifies the cover image to improve the robustness of data hiding technique. The cover image is transformed using Fractional Fourier Transform (FrFT) into a time-frequency domain and the optimal pixel locations for hiding the secret data are found using firefly algorithm. Firefly algorithm uses multi-objective function, which is a combination of Structural Similarity Index Measure (SSIM) and Bit Error rate (BER). The histogram shifting technique embeds secret data in the optimal pixel locations. The quality of test images is analyzed under varying payload as well as under varying fractional order. Experimental results conclude that this scheme produces good quality stego image. It has also been found from the simulation results that the proposed algorithm is more robust and reversible against various attacks as it provides lower bit error rate and higher normalization coefficient.

Evaluation of Machine Translation Quality through the Metrics of Error Rate and Accuracy

The aim of the paper is to find out whether it is necessary to use all automatic measures of error rate and accuracy when evaluating the quality of machine translation output from the synthetic Slovak language into the analytical English language. We used multiple comparisons for the analysis and visualized the results for each sentence through the icon graphs. Based on the results, we can state that all examined metrics, which are based on textual similarity, except the f-measure, are needed to be included in the MT quality evaluation when analyzing, sentence by sentence, the machine translation output.

Optimal design of shear vertical wave electromagnetic acoustic transducers in resonant mode

In this paper, a new electromagnetic acoustic resonance (EMAR) transducer is proposed for precise thickness measurement in specimen. The new EMAR is composed of a mirror symmetric coil (MSC) and a pair of Nd-Fe-B permanent magnets with the different polarity for enhancing the generation and detection of resonant signals. Firstly, a finite element model was established to simulate the distributions of Lorentz force produced by new EMAR and the resonant process of shear waves. Furthermore, the relationship between the frequency response characteristic of the new EMAR and the common EMAR were explored. Finally, to verify the performance of the EMAR, several experiments were performed. Compared with the common EMAR transducer, the resonant amplitude of the new EMAR transducer was increased by 121.74% and the signal-to-noise ratio was increased by 28.35%, and the resonance frequency interval of the new EMAR was twice that of the common mode in the frequency domain simulation experiment, this advantage effectively reduced the error rate of measurement. The results show that the new EMAR transducer with mirror coil structure has higher accuracy in thickness detection of specimens.