Error Rate Reduction Research Articles

Age estimation in humans through the analysis of aspartic acid racemization from teeth: A scoping review of methods, outcomes, and open research questions

Teeth are considered the most resistant structures in the human body. In forensic odontology, teeth are useful for human identification, especially when dental age estimation is necessary. Despite numerous studies, there is no consensus regarding the best methods for dental age estimation. The analysis of aspartic acid racemization, however, has shown promising results. This scoping review aimed to present a descriptive synthesis of the current literature regarding dental age estimation through aspartic acid racemization. Four electronic databases were screened: PubMed, Scielo, Web of Science, and Scopus. Cross-sectional studies published before April 2021 were selected. From 206 articles found, 26 met the eligibility criteria. Several experimental protocols and laboratory settings were detected, but the different protocols did not seem to significantly reduce error rates in dental age estimation. The analysis of aspartic acid racemization in human dental tissues produced accurate and potentially reliable results for age estimation. Aspartic acid racemization stands out especially in the adulthood – age category in which other methods struggle to deliver proper performances. Studies with larger samples, independent testing, and standardized laboratory procedures are necessary. Equator-like reporting guidelines are encouraged to enable future systematic reviews and meta-analyses.

An Efficient Handwritten Digit Recognition Based on Convolutional Neural Networks with Orthogonal Learning Strategies

The predictions of characters/text/digits from the handwritten images have made the research community spotlight towards recognition. There are enormous applications and ambiguity that made prediction possible with Deep Learning (DL) approaches. Primarily, there are four necessary steps to be carried out with handwriting prediction. First, consideration of a dataset that is more appropriate for DL validation an inefficient manner. Here, Special Database 1 and Special Database 2 are used, which are combined and modified by the National Institute of Standards and Technology (NIST). Next is pre-processing of input handwritten digit recognition data by data normalization, extraction of efficient features which provides better prediction accuracy. The proposed idea uses pixel values as features with the analysis of hyper-parameters to enhance near-human performance. With SVM, non-linear and linear models are built to extract the appropriate features for further processing. The features are separate and placed over the Bag of Features (BoF), which is used by the next processing stage. Finally, a novel Convolutional Neural Network (CNN) is by built modifying the network structure with Orthogonal Learning Particle Swarm Optimization (CNN-OLPSO). This modification is adopted for evolutionarily optimizing the number of hyper-parameters. This proposed optimizer predicts the optimal values from the fitness computation and shows better efficiency when compared to various other conventional approaches. The novelty which relies on CNN adoption is to endeavor a suitable path towards digitalization and preserve the handwritten structure and help automatic feature extraction using CNN by offering better computation accuracy. The optimization approach helps to avoid over-fitting and under-fitting issues. Here, metrics like accuracy, elapsed time, recall, precision, and [Formula: see text]-measure are evaluated. The results of CNN-OLPSO give better accuracy, reduced error rate and better execution time (s) compared to other existing methods. Thus, the proposed model shows better tradeoff in the recognition rate of handwritten digits.

IoT-assisted physical education training network virtualization and resource management using a deep reinforcement learning system

The Internet of Things (IoT) development made it possible for technology to communicate physical education by connecting cost-effective heterogeneous devices and digital applications to uncontrolled and accessible environments. The traditional physical education monitoring environment creates crucial manual efforts on athletes' activity observations and tracking consistently. Similarly, remote monitoring and assessment of athletes in sports training seem to be barriers to physical education monitoring and training. It creates various chances to improve training and education through technology advancements like IoT and deep learning. Students can efficiently monitor their physical behavior to increase their physical and psychological benefits. The IoT-assisted physical activity monitoring device is proposed to track students' physical activity and enhance outcomes. The management ability allows students to organize and increase speed their physical activity in a wellness manner. In addition, this study examines the connections between monitoring ability which is an essential component for sports activities and physical activity. This system collects essential information from IoT-based wearable devices that interact with the data in real time by virtualizing the device. The IoT network includes several device activities and monitors the heartbeat and physical body temperature of a person. The analysis of specific studies and student feedback shows that the designed virtual system of physical educations is effective in its application and implementation and provides a reliable guide for developing student physical educational systems. The experimental analysis is evaluated; the solution offered is developing and supporting physical education and training approaches in reality and creates healthy environment systems to solve the health monitoring challenges posed by IoT devices. The proposed method has achieved extraordinary physical activity monitoring compared to the conventional systems, as shown by experimental findings. The simulation analysis of physical education can help students and improve the associated aspects of physical abilities with high accuracy ratio (98.3), prediction ratio (96.5%), interaction ratio (94.4%), performance ratio (95.1%), the efficiency ratio (93.2), F-score (92.2%), and reduce error rate (17.5%) and physical activity patterns.

Low‐latency transformer model for streaming automatic speech recognition

Transformer models have made great progress in automatic speech recognition. However, it is challenging for streaming transformer models to make trade-off between output latency and recognition accuracy. In this letter, it is aimed to propose a low-latency transformer model with satisfactory recognition accuracy. First, a streaming transformer is designed and explain how it works streamingly. Second, the authors propose to use CTC during training to minimise the latency of transformer models. Finally, the authors also propose to utilise CTC as a backup during decoding to ensure that the low-latency characteristic is maintained. The authors fairly compare our streaming transformer model to existing streaming models, particularly the transducer model, which is a popular low-latency approach. The experiments show that, while having comparable output latency, the transformer model outperforms the transducer model by average relative character (or word) error rate reduction of 22.18%, 26.71% and 19.36% on HKUST, Switchboard and Call Home, respectively.

EP.FRI.376 Early Routine CT Scans for All Surgical Admissions for Abdominal Pain May Improve Patient Outcomes

Abstract Aim Acute abdominal pain remains a diagnostic challenge in the Emergency Department (ED) as pathologies can involve various surgical craft groups. Computed tomography (CT) enables accurate diagnosis of abdominal pathologies. However, with pressures on ED such as the “4 hour rule” established by the Australian NSW Health Emergency Performance Plan, there may be resistance or omission of early CT in ED. We hypothesise that early, routine CT in adult patients presenting with acute abdominal pain requiring hospital admission improves patient outcomes by increasing diagnostic and referral accuracy. This study compares the proportion of correct ED diagnosis of abdominal pain presentations with and without formal imaging reports. Methods Data from 118 patients presenting with abdominal pain are collected prospectively in a regional hospital and analysed. Patient demographics, imaging results, initial ED diagnosis and final discharge diagnosis are further examined. Results Out of the 118 patients who had abdominal pain, 32 patients obtained complete imaging with a radiology report whilst 86 patients either did not have any imaging performed or was referred to a general surgical unit prior to obtaining a formal report. Among the patients who had imaging reported, 78% (n = 25/32) had the correct diagnosis, whilst those without a radiology report had a 52% (n = 45/86) diagnostic accuracy. This demonstrates an improved accuracy of diagnosis or reduced error rate of 26% when a scan report is available (p = 0.01). Conclusion Early, routine CT with formal reporting significantly reduces diagnostic error rates and increases accurate referral. This allows accurate diagnosis and improves patient outcomes.

Spin-transfer torque switching probability of CoFeB/MgO/CoFeB magnetic tunnel junctions beyond macrospin

We present an empirical description of experimental spin-torque switching probability for the CoFeB/MgO/CoFeB type of magnetic tunnel junctions beyond macrospin limit, parametrizing measurement data for direct comparison with the corresponding macrospin asymptote expression. We show that, near 35 nm in diameter, spin-torque switching speed in these tunnel devices is faster than macrospin-limit predictions. These devices have a faster reduction of switching error rate versus spin-torque drive amplitude than macrospin. While the functional form similar to macrospin can still describe experimental data satisfactorily, the parameters no longer correspond to materials values. Instead they reflect the nonuniform nature of the switching process. Further, the parameters depend on the resistance-area product ${r}_{A}$ of the junction, with higher ${r}_{A}$ causing a steeper slope of switching error versus switching current. This ${r}_{A}$ dependence could not originate from low-bias spin-dependent tunneling. These observations suggest that, in addition to nonuniform nonlinear dynamics during switching, it is also important to consider higher-order dynamic processes, including a high-bias tunnel electron's spin-flip scattering, voltage-induced change to interface magnetism, and possibly Joule heating.

High-throughput hardware deployment of pruned neural network based nonlinear equalization for 100-Gbps short-reach optical interconnect.

Hardware implementation of neural network based nonlinear equalizers will encounter tremendous challenges due to a high-throughput data stream and high computational complexity for 100-Gbps short-reach optical interconnects. In this Letter, we propose a parallel pruned neural network equalizer for high-throughput signal processing and minimized hardware resources. The structure of a time-interleaved neural network equalizer with a delay module is deployed in a field programmable gate array with advanced pruned algorithms, demonstrating significant bit error rate reduction for 100-Gbps real-time throughput with 200 parallel channels. Moreover, the dependence of processing throughput, hardware resources, and equalization performance is investigated, showing that over 50% resource reduction without performance degradation can be achieved with the pruning strategy.

One- and Two-Phase Software Requirement Classification Using Ensemble Deep Learning.

Recently, deep learning (DL) has been utilized successfully in different fields, achieving remarkable results. Thus, there is a noticeable focus on DL approaches to automate software engineering (SE) tasks such as maintenance, requirement extraction, and classification. An advanced utilization of DL is the ensemble approach, which aims to reduce error rates and learning time and improve performance. In this research, three ensemble approaches were applied: accuracy as a weight ensemble, mean ensemble, and accuracy per class as a weight ensemble with a combination of four different DL models—long short-term memory (LSTM), bidirectional long short-term memory (BiLSTM), a gated recurrent unit (GRU), and a convolutional neural network (CNN)—in order to classify the software requirement (SR) specification, the binary classification of SRs into functional requirement (FRs) or non-functional requirements (NFRs), and the multi-label classification of both FRs and NFRs into further experimental classes. The models were trained and tested on the PROMISE dataset. A one-phase classification system was developed to classify SRs directly into one of the 17 multi-classes of FRs and NFRs. In addition, a two-phase classification system was developed to classify SRs first into FRs or NFRs and to pass the output to the second phase of multi-class classification to 17 classes. The experimental results demonstrated that the proposed classification systems can lead to a competitive classification performance compared to the state-of-the-art methods. The two-phase classification system proved its robustness against the one-phase classification system, as it obtained a 95.7% accuracy in the binary classification phase and a 93.4% accuracy in the second phase of NFR and FR multi-class classification.

Mobility speed prediction using ARIMA and RNN for random walk mobility model in mobile ad hoc networks

AbstractIn this article, an auto‐regressive integrated moving average (ARIMA) modeling has been proposed to predict the mobility speed of the nodes in the mobile ad hoc networks (MANETs). The mobility speed prediction supports effective route discovery to enhance efficient and reliable routing. The random walk mobility model had been used to forecast the mobility of the nodes. Subsequently, various the recurrent neural network (RNN) model has been developed with different number hidden neurons, and the prediction results are compared with proposed ARIMA based results. The Akaike information criterion, auto‐correlation function metrics are evaluated to assess the dataset's quality. Several network scenarios are evaluated with various speeds and the number of nodes. The results of the proposed ARIMA model are compared with the RNN results. It demonstrates that the proposed model has higher prediction rates with reduced error rates of 0.1%–1.4% than RNN values. Such that the proposed methods facilitate the higher prediction rate of 10%–26% as compared with the RNN. Therefore, the proposed mobility speed prediction techniques support the effective routing in the MANETs.

Confidence Learning for Semi-Supervised Acoustic Event Detection

In recent years, the involvement of synthetic strongly labeled data, weakly labeled data, and unlabeled data has drawn much research attention in semi-supervised acoustic event detection (SAED). The classic self-training method carries out predictions for unlabeled data and then selects predictions with high probabilities as pseudo-labels for retraining. Such models have shown its effectiveness in SAED. However, probabilities are poorly calibrated confidence estimates, and samples with low probabilities are ignored. Hence, we introduce a confidence-based semi-supervised Acoustic event detection (C-SAED) framework. The C-SAED method learns confidence deliberately and retrains all data distinctly by applying confidence as weights. Additionally, we apply a power pooling function whose coefficient can be trained automatically and use weakly labeled data more efficiently. The experimental results demonstrate that the generated confidence is proportional to the accuracy of the predictions. Our C-SAED framework achieves a relative error rate reduction of 34% in contrast to the baseline model.

Intranasal vasopressin like oxytocin increases social attention by influencing top-down control, but additionally enhances bottom-up control

The respective roles of the neuropeptides arginine vasopressin (AVP) and oxytocin (OXT) in modulating social cognition and for therapeutic intervention in autism spectrum disorder have not been fully established. In particular, while numerous studies have demonstrated effects of oxytocin in promoting social attention the role of AVP has not been examined. The present study employed a randomized, double-blind, placebo (PLC)-controlled between-subject design to explore the social- and emotion-specific effects of AVP on both bottom-up and top-down attention processing with a validated emotional anti-saccade eye-tracking paradigm in 80 healthy male subjects (PLC = 40, AVP = 40). Our findings showed that AVP increased the error rate for social (angry, fearful, happy, neutral and sad faces) but not non-social (oval shapes) stimuli during the anti-saccade condition and reduced error rates in the pro-saccade condition. Comparison of these findings with a previous study (sample size: PLC = 33, OXT = 33) using intranasal oxytocin revealed similar effects of the two peptides on anti-saccade errors, although with some difference in effects of specific face emotions, but a significantly greater effect of AVP on pro-saccades. Both peptides also produced a post-task anxiolytic effect by reducing state anxiety. Together these findings suggested that both AVP and OXT decrease goal-directed top-down attention control to social salient stimuli but that AVP more potently increased bottom-up social attentional processing.

Automation of Educational Institution Documentation

The article addresses issues related to the implementation of the competence approach in order to make the teaching process of an educational institution more efficient, analyzes the reasons for diminished efficiency of putting into practice the automated systems used for university management. The authors propose the structure of an automated system, which allows the university to create the required package of teaching documents taking into account the three-level system: administrator, supervisor of educational program, and lecturer. Designing object-oriented model in the area of a special subject requires identification of business-requirements to the system. On the basis of the model of subject area and specified business-requirements, the diagram of applying variants and the schemes of functioning is constructed. The suggested system of the automated execution of teaching documents enables to reduce error rate while carrying out and implementing it due to the integrated system of conceptions and feedback between its elements.

Comparison of seven single cell whole genome amplification commercial kits using targeted sequencing

Advances in whole genome amplification (WGA) techniques enable understanding of the genomic sequence at a single cell level. Demand for single cell dedicated WGA kits (scWGA) has led to the development of several commercial kit. To this point, no robust comparison of all available kits was performed. Here, we benchmark an economical assay, comparing all commercially available scWGA kits. Our comparison is based on targeted sequencing of thousands of genomic loci, including highly mutable regions, from a large cohort of human single cells. Using this approach we have demonstrated the superiority of Ampli1 in genome coverage and of RepliG in reduced error rate. In summary, we show that no single kit is optimal across all categories, highlighting the need for a dedicated kit selection in accordance with experimental requirements.

Semi-supervised parallel shared encoders for speech emotion recognition

Supervised speech emotion recognition requires a large number of labeled samples that limit its use in practice. Due to easy access to unlabeled samples, a new semi-supervised method based on auto-encoders is proposed in this paper for speech emotion recognition. The proposed method performed the classification operation by extracting the information contained in unlabeled samples and combining it with the information in labeled samples. In addition, it employed maximum mean discrepancy cost function to reduce the distribution difference when the labeled and unlabeled samples were gathered from different datasets. Experimental results obtained on different emotional speech datasets demonstrated that the proposed method reached better performance than previous methods in matched, semi-matched, and mismatched conditions. As an example, the proposed method boosted the error reduction rate on the INTERSPEECH 2009 Emotion challenge task on average by 14.13% via utilizing only 200 labeled samples. Besides, the proposed method was investigated as a domain adaptation method for recognizing Persian emotional speech. In this case, the proposed method boosted the accuracy of recognition by 10% compared to that of the cross-training method when German emotional database was used as the source domain.

Accuracy of heart girth tapes in the estimation of weights of pre-weaned calves.

BackgroundHeart girth tapes (HGTs) are often used as an alternative to weight scales for calves. This study investigated the accuracy of HGT in estimating bodyweight and daily liveweight gain (DWLG) of pre‐weaned calves, and the impact of inter‐observer variation.MethodIn Study 1, 119 calves were weighed using HGT and electronic scales on multiple occasions. Mixed‐effects models for both bodyweight and DLWG were used to determine the accuracy of HGT compared to the electronic scales. Simulation data were used to further analyse the accuracy of DLWG estimation including for factors such as the effect of group size on group DLWG estimates.In Study 2, 10 observers weighed 20 pre‐weaned calves, using HGT and electronic scales. Mixed‐effect model was used to investigate the impact of different observers on the accuracy of HGT on measuring bodyweights.ResultsMixed‐effects model results suggest HGT provides a relatively accurate estimation of weight (MAE: 2.66 kg) and relatively inaccurate estimation of DLWG (MAE 0.10 kg/d). Simulated data identified associations between time between weight dates and error in DLWG estimation, with MAE of individual DLWG estimation decreasing from 0.43 kg/d when 14 days apart to 0.08 kg/d when 70 days apart. Increased calf numbers reduced error rates of group DLWG estimation, with <0.05 kg/d error achieved in >90% of simulations when 12 calves were weighed 70 days apart.ConclusionsHGTs are relatively accurate at estimating individual bodyweights but are unreliable methods for measuring DLWG in individual calves, particularly weighed within a short‐time period. Estimates at group level however are relatively accurate, providing there is a suitable period of time between weigh dates and an appropriate number of calves per group.

Modulation of Methamphetamine-Related Attention Bias by Intermittent Theta-Burst Stimulation on Left Dorsolateral Prefrontal Cortex.

BackgroundPrevious studies have identified the treatment effect of repetitive transcranial magnetic stimulation (rTMS) on cravings of patients with methamphetamine use disorder (MUD). However, the mechanism underlying the treatment effect remains largely unknown. A potential candidate mechanism could be that rTMS over the dorsolateral prefrontal cortex (DLPFC) modulates the attention bias to methamphetamine-related cues. The purpose of this study is therefore to determine the modulation of rTMS on methamphetamine-related attention bias and the corresponding electrophysiological changes.MethodsForty-nine patients with severe MUD were included for analysis. The subjects were randomized to receive the active intermittent theta-burst stimulation (iTBS) or sham iTBS targeting DLPFC for 20 sessions. Participants performed the Addiction Stroop Task before and after the treatment while being recorded by a 64-channel electroencephalogram. Baseline characteristics were collected through the Addiction Severity Index.ResultsPost-treatment evaluations showed a reduced error rate in discriminating the color of methamphetamine words in the active iTBS group compared with the sham iTBS group. Following rTMS treatment, we found the significant time-by-group effect for the N1 amplitude (methamphetamine words > neutral words) and P3 latency (methamphetamine words > neutral words). The change of N1 amplitude was positively correlated with cravings in the active group. Moreover, reduced power of neural oscillation in the beta band, manifesting at frontal central areas, was also found in the active group.ConclusionThis study suggests that attention bias and the beta oscillation during the attentional processing of methamphetamine words in patients with MUD could be modulated by iTBS applied to left DLPFC.

Improving convolutional neural networks performance for image classification using test time augmentation: a case study using MURA dataset.

Bone fractures are one of the main causes to visit the emergency room (ER); the primary method to detect bone fractures is using X-Ray images. X-Ray images require an experienced radiologist to classify them; however, an experienced radiologist is not always available in the ER. An accurate automatic X-Ray image classifier in the ER can help reduce error rates by providing an instant second opinion to the emergency doctor. Deep learning is an emerging trend in artificial intelligence, where an automatic classifier can be trained to classify musculoskeletal images. Image augmentations techniques have proven their usefulness in increasing the deep learning model's performance. Usually, in the image classification domain, the augmentation techniques are used during training the network and not during the testing phase. Test time augmentation (TTA) can increase the model prediction by providing, with a negligible computational cost, several transformations for the same image. In this paper, we investigated the effect of TTA on image classification performance on the MURA dataset. Nine different augmentation techniques were evaluated to determine their performance compared to predictions without TTA. Two ensemble techniques were assessed as well, the majority vote and the average vote. Based on our results, TTA increased classification performance significantly, especially for models with a low score.

Hardware-Efficient Leakage-Reduction Scheme for Quantum Error Correction with Superconducting Transmon Qubits

Leakage outside of the qubit computational subspace poses a threatening challenge to quantum error correction (QEC). We propose a scheme using two leakage-reduction units (LRUs) that mitigate these issues for a transmon-based surface code, without requiring an overhead in terms of hardware or QEC-cycle time as in previous proposals. For data qubits we consider a microwave drive to transfer leakage to the readout resonator, where it quickly decays, ensuring that this negligibly affects the coherence within the computational subspace for realistic system parameters. For ancilla qubits we apply a $|1\rangle\leftrightarrow|2\rangle$ $\pi$ pulse conditioned on the measurement outcome. Using density-matrix simulations of the distance-3 surface code we show that the average leakage lifetime is reduced to almost 1 QEC cycle, even when the LRUs are implemented with limited fidelity. Furthermore, we show that this leads to a significant reduction of the logical error rate. This LRU scheme opens the prospect for near-term scalable QEC demonstrations.

Autonomous Stair Climbing Robot

An autonomous robot, also known as simply an autorobot or autobot, is a robot that performs behaviors or tasks with a high degree of autonomy. Autonomous robotics is usually considered to be a subfield of artificial intelligence, robotics, and information engineering. Specifically, “autonomous robots can help: Increase efficiency and productivity. Reduce error, re-work, and risk rates. Improve safety for employees in high-risk work”environments. In this COVID era every 4 out of the 5 Fortune 500 e-commerce companies and especially the e-commerce giants like Walmart , Amazon and Flipkart are focusing even more on contact less delivery. Delivery using automated bots is a concept that can be used even after the Covid situation gets over. The popularity of e-commerce buying is exploding, and the phenomenon has huge implications for shippers and distributors. To paraphrase Charles Darwin, “those companies best prepared to adapt to this new reality are the ones that will flourish while laggards will be left in the dust”. “Logistics is all about movement from one place to another. This movement happens across facilities and within facilities. While AGVs (AUTOMATED GUIDED VEHICLES) have found their way into applications of material movement inside a warehouse, they are designed to move across flat surfaces with minor undulations (in some cases requiring specialised mezzanine surfaces). This paper comes up with conceptual building and designing an autonomous robot which can work on flat surfaces and even climb up and down standard stairs with a payload of 3-5kg. This paper further will inform you about the conceptual building of an autonomous stair climbing robot involving all the steps like sensor fusion , navigation , tracking , mapping and path planning.

Internet of vehicles and autonomous systems with AI for medical things.

The current world faces a considerable traffic rate on roads due to the increase in various types of vehicles. It caused emergency vehicles to delay and increasing the patients' health risk factor. Internet of vehicle-based artificial neural network (IoV-ANN) has been proposed for effective health autonomous system in medical things. The proposed IoV-ANN provides a secure network to monitor and track the vehicle's motion using the global positioning system. It consists of an autonomous system which is enabled with an artificial neural network (ANN). ANN model has three layers. First layers collect the data using IoV sensors. Second or hidden layers process the sensor data, predict the road's traffic condition and reroute the emergency vehicle to an exact route. IoV-ANN helps to reduce road congestion in this article to enhance the timely functioning of an emergency vehicle. ANN categorizes the congestion networks of traffic. Traffic restrictions such as changing the queue gap in the road signals and the alternative roads are altered automatically due to congestion. It allows the government to develop ideas for alternate routes to enhance traffic control. The output layer gives commands to the driver autonomously. The simulation analysis of the proposed method proved that the system could work independently. The IoV-ANN achieves the highest performance rate of (97.89%), with a reduced error rate (9.12%) traffic congestion rate (10.31%), travel period (32 s), vehicle detection rate (93.61%), classification accuracy (95.02%), MAPE (8.4%), throughput rate (93.50%) when compared to other popular methods.