Optimal Model Parameters Research Articles

Milk Yield Prediction and Economic Analysis of Optimized Rearing Environment in a Cold Region Using Neural Network Model

The milk yield of dairy cows in a non-stressed state in the cold region in China is lower during cold seasons. In this study, the correlations between indoor environmental factors and milk production were analyzed. Temperature, relative humidity, and light intensity were found to be the main factors affecting milk yield. The warning values of these factors for lower milk production were 5 °C, 60%, and 300 lx, respectively. A neural network model predicting milk yield based on environmental factors was established, and the optimal model parameters were determined, resulting in a high accuracy of R2 = 0.802. This model was used to investigate the optimal measure for improving the indoor environment, which helps to increase milk production and economic benefits, including LED lights, heating radiators, and dehumidifiers. In conclusion, each type of device led to the growth of milk yield, reaching 2.341, 1.706, and 1.893 kg cow−1 Day−1. The combination of heating radiator and LED light resulted in the highest increased net benefit of 16.802 CNY cow−1 Day−1. This is the first time that a neural network model was successfully built to predict milk yield based on climatic features which was also applied to economic analysis of indoor environment improvement for dairy barns in extreme cold regions.

Prediction technique for flow boiling heat transfer and critical heat flux in both microgravity and Earth gravity via artificial neural networks (ANNs)

This study is part of the Flow Boiling and Condensation Experiment (FBCE) and utilizes flow boiling data collected in both microgravity onboard the International Space Station (ISS) and Earth gravity at different channel orientations. The goal is to develop a prediction technique for heat transfer and critical heat flux (CHF) for flow boiling in both microgravity and Earth gravity using artificial neural networks (ANNs). The working fluid, n-perfluorohexane or FC-72, flows through a rectangular channel of 114.6 mm heated length, 2.5 mm heated width, and 5.0 mm unheated height with either one or two walls heated. The consolidated FBCE database for heat transfer coefficient comprises 29,226 datapoints spanning a mass velocity of 173 – 3200 kg/m2s, pressure of 102 – 238 kPa, subcooling of 0 – 44°C, and thermodynamic equilibrium quality of -0.60 – 0.95 (spanning from highly subcooled to high-quality saturated boiling). Following a statistical analysis of various input parameters relevant to flow boiling and optimization of key model parameters, a fully connected feed-forward ANN is developed to predict Nutp. It predicts the entire test database with an overall mean absolute error (MAE) of just 7.99% with consistent and accurate predictions in each subset. Similarly, 641 CHF datapoints from FBCE were consolidated into a database spanning a mass velocity of 99 – 3212 kg/m2s, inlet pressure of 97 – 239 kPa, inlet subcooling of 0 – 46°C, inlet thermodynamic equilibrium quality of -0.61 – 0.86, and CHF values of 4 – 54 W/cm2. A separate ANN is developed by following the same methodology as heat transfer, and it predicts dimensionless CHF, BoCHF, with an overall MAE of just 12.05%. Existing seminal correlations are assessed for subsets of the two consolidated FBCE databases, and the ANNs are shown to have better accuracies in each subset of the database. The ANNs’ high prediction accuracy, in conjunction with their ability to predict physical parametric trends in previously unseen data, shows their potential as prediction tools for both heat transfer and CHF for flow boiling in microgravity and Earth gravity.

Enhancing Skeletal Age Estimation Accuracy Using Support Vector Regression Models

ObjectiveThe objective of the study was to determine if support vector regression (SVR) models could enhance the accuracy of skeletal age estimation compared to original metrics. MethodThe study used a dataset of 5,018 individuals from Wuhan, spanning ages 1 to 17. Optimal model parameters were found using cross-validation and grid search techniques. The study compared SVR-based bone age assessment metrics with original metrics and evaluated the performance of the SVR model across different sample sizes. ResultsThe findings unequivocally demonstrated SVR's superior reliability over original metrics in assessing bone age among children in central China. Notably, the SVR-based metric using the combined TW3, CHN05 and GP yields the second highest accuracy, while the SVR-based metric using the combined TW3 and CHN05 yields the highest accuracy. ConclusionThis research highlights SVR's potential for accuracy improvement and robustness with limited datasets.

Particle Swarm Optimization for Penalize cox models in long-term prediction of breast cancer data

The particle swarm optimization algorithm (PSO) was applied to penalize the Cox model for predicting long-term outcomes of breast cancer patients. This study makes use of data on 198 patients' breast cancer survival, including their age, estrogen receptor status, tumor size, and grade, as well as the expression levels of 76 genes. The objective was to identify a subset of features that could accurately predict patient survival while controlling for overfitting and model complexity. PSO was used to search for optimal model parameters. The algorithm was designed to minimize a penalized partial likelihood function, which balances the tradeoff between an accurate model and model complexity. The values of the objective function were compared with other feature selection techniques, including the Least Absolute Shrinkage and Selection Operator (LASSO) and Elastic regression, and were found to perform better in relation to predictive accuracy and feature selection. The study results showed that PSO-based Cox models with cross-validation to regularization parameters had higher prediction accuracy than models trained with other feature selection methods. The PSO algorithm identified a subset of features that were consistently selected across multiple iterations, indicating their importance in predicting patient survival. Overall, the study demonstrates the potential of PSO-based feature selection in improving the accuracy and interpretability of Cox regression models for predicting long-term outcomes in breast cancer patients.

Improved detection of aberrant splicing with FRASER 2.0 and the intron Jaccard index

Detection of aberrantly spliced genes is an important step in RNA-seq-based rare-disease diagnostics. We recently developed FRASER, a denoising autoencoder-based method that outperformed alternative methods of detecting aberrant splicing. However, because FRASER's three splice metrics are partially redundant and tend to be sensitive to sequencing depth, we introduce here a more robust intron-excision metric, the intron Jaccard index, that combines the alternative donor, alternative acceptor, and intron-retention signal into a single value. Moreover, we optimized model parameters and filter cutoffs by using candidate rare-splice-disrupting variants as independent evidence. On 16,213 GTEx samples, our improved algorithm, FRASER 2.0, called typically 10 times fewer splicing outliers while increasing the proportion of candidate rare-splice-disrupting variants by 10-fold and substantially decreasing the effect of sequencing depth on the number of reported outliers. To lower the multiple-testing correction burden, we introduce an option to select the genes to be tested for each sample instead of a transcriptome-wide approach. This option can be particularly useful when prior information, such as candidate variants or genes, is available. Application on 303 rare-disease samples confirmed the relative reduction in the number of outlier calls for a slight loss of sensitivity; FRASER 2.0 recovered 22 out of 26 previously identified pathogenic splicing cases with default cutoffs and 24 when multiple-testing correction was limited to OMIM genes containing rare variants. Altogether, these methodological improvements contribute to more effective RNA-seq-based rare diagnostics by drastically reducing the amount of splicing outlier calls per sample at minimal loss of sensitivity.

Enhancing second-order blind identification for underdetermined operational modal analysis through bandlimited source separation

Blind source separation (BSS) has been demonstrated to be a promising method for operational modal analysis (OMA). However, the conventional BSS-based modal identification methods can only handle the determined or overdetermined identification where the number of sensors must equal or exceed that of active modes, limiting their wider application. Furthermore, the existing methods for underdetermined BSS require sparsity assumption of sources, optimization of model parameters, or structural prior information, undermining the non-parametric and “blind” properties of BSS. This paper proposes a novel method for underdetermined OMA. Firstly, the bandlimited source separation using sinc-dictionaries is performed to efficiently transform the underdetermined problem into several determined or overdetermined ones. An improved second-order blind identification (SOBI) involving Hilbert transform and random time-lagged covariances is then applied for each bandlimited signal to recover modal responses and mode shapes. To attenuate noise, an adaptive singular spectrum analysis (SSA) can be optionally applied to the recovered modal responses for the highly noisy systems. Finally, the natural frequencies and damping ratios are estimated by the single-degree-of-freedom (SDOF) fitting technique. Several numerical examples, an experimental frame case, and a field test of a pedestrian bridge are studied to verify the effectiveness of the proposed strategy.

A few-shot disease diagnosis decision making model based on meta-learning for general practice

BackgroundDiagnostic errors have become the biggest threat to the safety of patients in primary health care. General practitioners, as the “gatekeepers” of primary health care, have a responsibility to accurately diagnose patients. However, many general practitioners have insufficient knowledge and clinical experience in some diseases. Clinical decision making tools need to be developed to effectively improve the diagnostic process in primary health care. The long-tailed class distributions of medical datasets are challenging for many popular decision making models based on deep learning, which have difficulty predicting few-shot diseases. Meta-learning is a new strategy for solving few-shot problems. Methods and materialsIn this study, a few-shot disease diagnosis decision making model based on a model-agnostic meta-learning algorithm (FSDD-MAML) is proposed. The MAML algorithm is applied in a knowledge graph-based disease diagnosis model to find the optimal model parameters. Moreover, FSDD-MAML can learn learning rates for all modules of the knowledge graph-based disease diagnosis model. For n-way, k-shot learning tasks, the inner loop of FSDD-MAML performs multiple gradient update steps to learn internal features in disease classification tasks using n×k examples, and the outer loop of FSDD-MAML optimizes the meta-objective to find the associated optimal parameters and learning rates. FSDD-MAML is compared with the original knowledge graph-based disease diagnosis model and other meta-learning algorithms based on an abdominal disease dataset. ResultMeta-learning algorithms can greatly improve the performance of models in top-1 evaluation compared with top-3, top-5, and top-10 evaluations. The proposed decision making model FSDD-MAML outperforms all the other models, with a precision@1 of 90.02 %. We achieve state-of-the-art performance in the diagnosis of all diseases, and the prediction performance for few-shot diseases is greatly improved. For the two groups with the fewest examples of diseases, FSDD-MAML achieves relative increases in precision@1 of 29.13 % and 21.63 % compared with the original knowledge graph-based disease diagnosis model. In addition, we analyze the reasoning process of several few-shot disease predictions and provide an explanation for the results. ConclusionThe decision making model based on meta-learning proposed in this paper can support the rapid diagnosis of diseases in general practice and is especially capable of helping general practitioners diagnose few-shot diseases. This study is of profound significance for the exploration and application of meta-learning to few-shot disease assessment in general practice.

Reliable and precise determination of energy conversion in fuel cells using an integrable energy model

Fuel cells serve as devices that directly convert chemical energy into electricity via electrochemical reactions. Assessing and characterizing fuel cell performance typically involves the creation of polarization curves. Nevertheless, currently employed polarization curve models have shown limited effectiveness when it comes to evaluating the conversion of chemical energy into electrical energy. To address this knowledge gap, this study has developed a concise and integrable energy model of fuel cells. This model aims to predict output voltage with a high degree of reliability and precision, relying on current density. The validity of this model was through experimentation, involving a dataset of twenty results. Fitting outcomes from these experiments demonstrated the capacity of the proposed model to faithfully replicate experimental curves. Furthermore, this model permits the dependable and precise calculation of fuel cell energy density, contingent upon optimal model parameters. In addition, this study has established quantitative correlations between the inherent characteristics of fuel cells and their energy density. Lastly, the proposed model facilitates the interpolation and extrapolation of polarization curves that may not be attainable through direct experimentation. This capability is achieved through the utilization of a secondary model that quantitatively associates the intrinsic attributes of fuel cells with optimal model coefficients.

Application of the combination on data-driven and discrete element model: Particle segregation prediction and classification in rotary drums

With the complexity and variability of particle systems in industrial scenarios and the huge cost of repetitive testing, predicting and classifying particle systems in advance have a certain necessity. Combined with discrete element method (DEM) and data-driven model, this work develops a prediction and classification model for particle flow characteristics based on the long-short term memory (LSTM) method. On the basis of the generated macroscopic and microscopic particle flow characteristic information, the LSTM model was constructed and then described particle flow information based on optimal model parameters determined by hyperparameter optimization. Compared of predicted results with the DEM simulation validated the interpolation and extrapolation ability of the established model. And it has better prediction accuracy and performance comparing with traditional prediction methods. In addition, for the classification of multiple working condition categories, the LSTM model trained from fused dataset obtained the best results with the accuracy of more than 95 % in all cases. With the combination of DEM model and data-driven approach, the rapid prediction of particle flow characteristics and parameter optimization in various industrial applications will be achieved.

A parameter optimization method of multi-layer nested slit resonator for the converter transformer noise in high voltage direct current converter station

We propose a parameter optimization method of multi-layer nested slit resonator (MNSR) for the converter transformer noise in a high voltage direct current (HVDC) converter station. With the optimization objective of achieving quasi-perfect sound absorption at the target frequencies and the structural parameters of MNSR as the optimization variables, the optimization model of the structural parameters of MNSR is established and then solved by the sequential quadratic programming algorithm (SQP). The theoretical and simulation results of the four validation cases show that the optimized MNSR can achieve quasi-perfect sound absorption at all target frequencies, which is consistent with the optimization objective. The parameter optimization design method provides an efficient method to design a dual-frequency quasi-perfect sound absorber for the converter transformer noise.

Design and Test of a Comb-Brush-Type Honeysuckle-Picking Device

In response to the issues of high honeysuckle-picking costs and low efficiency in honeysuckle picking, this study has devised a comb-brush-type picking device, considering the unique characteristics of honeysuckle plants. We elucidated the device’s structure and operational principles and designed critical components within the picking mechanism. Subsequently, through theoretical analysis, we identified the primary factors influencing the device’s operational performance. We then used the honeysuckle picking rates, honeysuckle breakage rates, and impurity rates as assessment metrics. Utilizing a one-factor test, we determined the permissible ranges for each factor. Employing the response surface methodology, we analyzed the interactions among these factors and conducted model parameter optimization. This optimization identified the optimal parameter combination: a forward speed of 3.99 km/h, a driving shaft speed of 316.53 rpm, and a picking teeth length of 70 mm. Finally, we performed verification tests using these optimized parameters. The results demonstrated that the maximum relative error between test verification values and model-optimized predictions was 4.86%. This outcome confirms that the comb-brush-type honeysuckle-picking device can meet the operational requirements of mechanized harvesting and offers valuable insights for developing harvesting devices for vine plants.

Online interoperable resources for building hippocampal neuron models via the Hippocampus Hub.

To build biophysically detailed models of brain cells, circuits, and regions, a data-driven approach is increasingly being adopted. This helps to obtain a simulated activity that reproduces the experimentally recorded neural dynamics as faithfully as possible, and to turn the model into a useful framework for making predictions based on the principles governing the nature of neural cells. In such a context, the access to existing neural models and data outstandingly facilitates the work of computational neuroscientists and fosters its novelty, as the scientific community grows wider and neural models progressively increase in type, size, and number. Nonetheless, even when accessibility is guaranteed, data and models are rarely reused since it is difficult to retrieve, extract and/or understand relevant information and scientists are often required to download and modify individual files, perform neural data analysis, optimize model parameters, and run simulations, on their own and with their own resources. While focusing on the construction of biophysically and morphologically accurate models of hippocampal cells, we have created an online resource, the Build section of the Hippocampus Hub -a scientific portal for research on the hippocampus- that gathers data and models from different online open repositories and allows their collection as the first step of a single cell model building workflow. Interoperability of tools and data is the key feature of the work we are presenting. Through a simple click-and-collect procedure, like filling the shopping cart of an online store, researchers can intuitively select the files of interest (i.e., electrophysiological recordings, neural morphology, and model components), and get started with the construction of a data-driven hippocampal neuron model. Such a workflow importantly includes a model optimization process, which leverages high performance computing resources transparently granted to the users, and a framework for running simulations of the optimized model, both available through the EBRAINS Hodgkin-Huxley Neuron Builder online tool.

An optimization model of tractor clutch pedal design parameters based on lower limb biomechanical characteristics

The design of tractor pedal affects operator's comfort, working efficiency, and occupational health. To optimize tractor clutch pedal design parameters, we conducted a single-factor experiment and a response surface experiment based on a biomechanical model. The experimental indicators included the biomechanical loads of major activated muscles and joints of the lower limb, and the experimental factors comprised pedal-seat horizontal distance L, pedal-seat vertical distance H, and pedal spring stiffness k. The results indicate that L and H mainly influence the loads of semitendinosus and hip. Semitendinosus activity and hip torque first decreased and then increased with increasing L (or H), reaching the lowest level at 82 cm of L (or 42 cm of H). The increase in k led to gradual decreases in iliopsoas activity and increases in semitendinosus, rectus femoris, soleus activities, and ankle torque, as well as a first decreasing and then increasing trend of hip and knee torques. The overall biomechanical loads of the lower limb first declined and then grew with increasing L, H, and k. The optimal range of clutch pedal design parameters was calculated to be {(L, H, k) | 36≤L ≤ 40, 72≤H ≤ 76, 14≤k ≤ 22} when using the response surface method. Within this range, muscle activity and joint torque were reduced by 14.2% and 51.6%, respectively, indicating effective alleviation of the lower limb biomechanical loads. This study provides important implications for the biomechanical analysis of tractor operator-pedal interaction system and lays a foundation for improving the design and comfort of tractors.

Fault Diagnosis of Analog Circuits Based on The Sailfish Algorithm Optimized SKELM

In view of the long training time for the optimization of the network model parameters of the SELM and the uncertainty of the model generalization ability, this paper proposes an analog circuit fault diagnosis model based on the sailfish algorithm to optimize the stacked kernel extreme learning machine(SKELM). This model introduces a kernel function to build a multi-layer KELM, which can improve the generalization ability and learning speed of the feedforward neural network. The weights of each layer of SKELM are obtained through the automatic encoder training based on the KELM. Since KELM-AE does not need to set initial parameters, the training speed is improved. However, the kernel parameters and regularization coefficients of KELM-AE are set manually, so the sailfish optimizer (SFO) is used to optimize these two parameters, and then the optimal SKELM model is built through layer by layer training. Finally, the Leap frog filter circuit is used as the simulation experiment circuit, and further compared with the optimized SELM. The results show that KELM-AE has strong generalization ability, and it can map fault features to high-dimensional feature space through nonlinear mapping without extracting fault features separately, thus improving the classification accuracy.

Multiscenarios Parameter Optimization Method for Active Disturbance Rejection Control of PMSM Based on Deep Reinforcement Learning

In this article, a multi-scenarios parameter optimization method for active disturbance rejection controller (ADRC) of permanent magnet synchronous motor (PMSM) based on deep reinforcement learning (DRL) is proposed as MSPO-DRL. The parameter setting of nonlinear ADRC has always been one of the difficulties affecting the optimal performance of ADRC, and there will be different optimal parameters under different control requirements. In this article, artificial intelligence algorithm is introduced into the parameter optimization process of ADRC, and the DRL parameter optimization model which can automatically optimize and adjust the ADRC parameters in different application scenarios is constructed so that the ADRC can achieve the best control effect conveniently, and the limitation of current methods has been solved. ADRC is applied to the speed loop of flux weakening control of PMSM for more electric aircraft, and the mathematical model of ADRC in this environment is built above all. The Markov decision process is integrated into ADRC. The interface module and reward function between ADRC and MSPO-DRL are designed. The concept of ADRC control scenario is defined and integrated into the concept of Markov decision process to improve the generalization of DRL. Then the MSPO-DRL model is established, and the deep deterministic gradient strategy is used as the gradient descent strategy to converge the parameters optimization. After the model learning is completed, different environmental conditions are randomly selected for simulation and experiments to verify the optimization effect and generalization performance of the algorithm. Optimizations which are carried out by the heuristic algorithms are used for comparisons, and the superiority and feasibility of the proposed algorithm is verified.

Predicting biochemical oxygen demand in wastewater treatment plant using advance extreme learning machine optimized by Bat algorithm

Wastewater quality modelling plays a vital role in planning and management of wastewater treatment plants (WWTP). This paper develops a new hybrid machine learning model based on extreme learning machine (ELM) optimized by Bat algorithm (ELM-Bat) for modelling five day effluent biochemical oxygen demand (BOD5). Specifically, this hybrid model combines the Bat algorithm for model parameters optimization and the standalone ELM. The proposed model was developed using historical measured effluents wastewater quality variables, i.e., the chemical oxygen demand (COD), temperature, pH, total suspended solid (TSS), specific conductance (SC) and the wastewater flow (Q). The performances of the hybrid ELM-Bat were compared with those of the multilayer perceptron neural network (MLPNN), the random forest regression (RFR), the Gaussian process regression (GPR), the random vector functional link network (RVFL), and the multiple linear regression (MLR) models. By comparing several input variables combination, the improvement achieved in the accuracy of prediction through the hybrid ELM-Bat was quantified. All models were first calibrated using training dataset and later tested using validation and based on four performances metrics namely, root mean square error (RMSE), mean absolute error (MAE), the correlation coefficient (R), and the Nash-Sutcliffe model efficiency (NSE). In all, it is concluded that the ELM-Bat is the most accurate model when all the six input were included as input variables, and it outperforms all other benchmark models in terms of predictive accuracy, exhibiting RMSE, MAE, R and NSE values of approximately, 0.885, 0.781, 2.621, and 1.989, respectively.

Machine Learning Techniques to Predict Timeliness of Care among Lung Cancer Patients.

Delays in the assessment, management, and treatment of lung cancer patients may adversely impact prognosis and survival. This study is the first to use machine learning techniques to predict the quality and timeliness of care among lung cancer patients, utilising data from the Victorian Lung Cancer Registry (VLCR) between 2011 and 2022, in Victoria, Australia. Predictor variables included demographic, clinical, hospital, and geographical socio-economic indices. Machine learning methods such as random forests, k-nearest neighbour, neural networks, and support vector machines were implemented and evaluated using 20% out-of-sample cross validations via the area under the curve (AUC). Optimal model parameters were selected based on 10-fold cross validation. There were 11,602 patients included in the analysis. Evaluated quality indicators included, primarily, overall proportion achieving "time from referral date to diagnosis date ≤ 28 days" and proportion achieving "time from diagnosis date to first treatment date (any intent) ≤ 14 days". Results showed that the support vector machine learning methods performed well, followed by nearest neighbour, based on out-of-sample AUCs of 0.89 (in-sample = 0.99) and 0.85 (in-sample = 0.99) for the first indicator, respectively. These models can be implemented in the registry databases to help healthcare workers identify patients who may not meet these indicators prospectively and enable timely interventions.

Quantitative Analysis of Biodiesel Adulterants Using Raman Spectroscopy Combined with Synergy Interval Partial Least Squares (siPLS) Algorithms

Biodiesel has emerged as an alternative to traditional fuels with the aim of reducing the impact on the environment. It is produced by the esterification of oleaginous seeds, animal fats, etc., with short-chain alcohols in an alkaline solution, which is one of the most commonly used methods. This increases the oxygen content (from the fatty acids) and promotes the fuel to burn faster and more efficiently. The accurate quantification of biodiesel is of paramount importance to the fuel market due to the possibility of adulteration, which can result in economic losses, engine performance issues and environmental concerns related to corrosion. In response to achieving this goal, in this work, synergy interval partial least squares (siPLS) algorithms in combination with Raman spectroscopy are used for the quantification of the biodiesel content. Different pretreatment methods are discussed to eliminate a large amount of redundant information of the original spectrum. The siPLS technique for extracting feature variables is then used to optimize the input variables after pretreatment, in order to enhance the predictive performance of the calibration model. Finally, the D1-MSC-siPLS calibration model is constructed based on the preprocessed spectra, the selected input variables and the optimized model parameters. Compared with the feature variable selection methods of interval partial least squares (iPLS) and backward interval partial least squares (biPLS), results elucidate that the D1-MSC-siPLS calibration model is superior to the D1-MSC-biPLS and the D1-MSC-iPLS in the quantitative analysis of adulterated biodiesel. The D1-MSC-siPLS calibration model demonstrates better predictive performance compared to the full spectrum PLS model, with the optimal determination coefficient of prediction (R2P) being 0.9899; the mean relative error of prediction (MREP) decreased from 9.51% to 6.31% and the root--mean-squared error of prediction (RMSEP) decreased from 0.1912% (v/v) to 0.1367% (v/v), respectively. The above results indicate that Raman spectroscopy combined with the D1-MSC-siPLS calibration model is a feasible method for the quantitative analysis of biodiesel in adulterated hybrid fuels.

Towards job screening and personality traits estimation from video transcriptions

In recent years, natural language processing (NLP) has gained new territory beyond its traditional use in text mining applications. This paper shows the effectiveness of NLP techniques in assessing the apparent human personality from his/her video transcript, building a bridge between NLP and computer vision-based reasoning. In this paper, a new deep learning model using attention mechanism and bidirectional LSTM layers for estimating the Big-five personality traits is provided and then tested on ChatLearn video dataset. The robustness of the approach is then tested by generalizing the method to two other datasets (B5 corpus and Mypersonality datasets). Several empirical evaluations taking into account the various inputs of the data processing pipeline have been performed to yield optimal model parameters. The developed model is tested on the APA’2016 competition dataset from Chalearn V2 Challenge Workshop for both Big-five personality trait estimation and interview score estimation. We achieved an average of 89.27% in personality trait recognition rate and 89.16% score in the Job Interview challenge. Similar trends of high accuracy score estimation is held for B5-corpus and MyPersonality datasets as well. The results outperformed several state-of-art approaches, demonstrating our approach’s feasibility in extending the computer vision approach of first impression personality to natural language processing. This opens up a new direction in multimedia analysis.

A phenomenological model of whole brain dynamics using a network of neural oscillators with power-coupling

We present a general, trainable oscillatory neural network as a large-scale model of brain dynamics. The model has a cascade of two stages - an oscillatory stage and a complex-valued feedforward stage - for modelling the relationship between structural connectivity and functional connectivity from neuroimaging data under resting brain conditions. Earlier works of large-scale brain dynamics that used Hopf oscillators used linear coupling of oscillators. A distinctive feature of the proposed model employs a novel form of coupling known as power coupling. Oscillatory networks based on power coupling can accurately model arbitrary multi-dimensional signals. Training the lateral connections in the oscillator layer is done by a modified form of Hebbian learning, whereas a variation of the complex backpropagation algorithm does training in the second stage. The proposed model can not only model the empirical functional connectivity with remarkable accuracy (correlation coefficient between simulated and empirical functional connectivity- 0.99) but also identify default mode network regions. In addition, we also inspected how structural loss in the brain can cause significant aberration in simulated functional connectivity and functional connectivity dynamics; and how it can be restored with optimized model parameters by an in silico perturbational study.