Efficient Computational Model Research Articles

An alternative biomimetic tool – Dual hydrophilic/reversed-phase interaction mode

During the early stages of drug discovery, it becomes increasingly important to acquire a basic physico-chemical and pharmacokinetic profile of drug candidates using efficient and cost-effective laboratory and computational models. In this study, the possibility of applying the dual hydrophilic interaction liquid chromatography/reversed-phase mode (HILIC/RP) as a alternative biomimetic tool for the preliminary profiling of imidazoline and piperazine-related compounds was investigated. Their retention behaviour on the mixed-mode HILIC/RP column was described as a dependence of the logarithm of the retention factor (logk) in relation to the volume fraction of the water component in the mobile phase (φ). The retention was quantified by integration of HILIC (logkH), RP (logkR) and the total HILIC/RP (logkA) area under the logk = f(φ) curve. The biomimetic capabilities of HILIC/RP mode were observed through physico-chemical and pharmacokinetic interpretation of chromatographic behaviour. The physico-chemical characteristics of the HILIC/RP mode were defined based on linear solvation energy relationships, genetic algorithm-multiple linear regression and Boruta-XGBoost predictive modeling. The created models showed reliable predictive performances for defining the logkA, logkR, and logkH values. The mixed-mode logkA, logkR, and logkH parameters correlate well with physico-chemical (e.g., lipophilicity indices) and basic pharmacokinetic properties of the examined compounds. The models created have confirmed that the simultaneous contribution of both HILIC/RP mechanisms can be used to accurately evaluate the key properties of newly synthesized imidazoline and piperazine-related compounds while reducing costs and selecting suitable candidates in the early stages of drug development and discovery.

Automated sentimental analysis using heuristic-based CNN-BiLSTM for E-commerce dataset

The sentimental analysis is processed according to structural detection, extraction, quantification, and evaluation of impacts and information present in natural language processing, biometric data, computer-language society, and text analysis. Advanced implementations in data analytics occur, which allows for identifying the underlining trends based on certain efficient computational models. Some of the social and e-commerce platforms consist of numerous reviews regarding the online products that are posted by the end-users. This helps the developers to understand priceless insight during the process of product design. The most significant aim of this work is to develop a Heuristic-based sentimental analysis for E-Commerce data. Here, the pre-processing of input E-Commerce data is performed by blank space removal, stop word removal, and stemming. Here, the statistical features are extracted using Cross Similarity Score (CSS) and Joint Similarity Score (JSS) regarding “positive, negative, and neutral keywords”. In addition, the word2vec features are also extracted. From the two sets of features, “Convolutional Neural Network (CNN) with Bidirectional Long Short Term Memory (BiLSTM)” is separately used that is named as Optimal Hybrid CNN-BiLSTM (OH-CNN-BiLSTM), and performed the hybridization of two models for the sentimental analysis. The designed CNN-BiLSTM technique is ensured better outcomes when processing long text as it provides advantages through the ability of CNN to extract the features and superiority of Bi-LSTM in learning the bidirectional dependencies of the long-term text. The main novelty of this work is to develop an improved Galactic Swarm Optimization (IGSO) algorithm for improving the hybridized model, as it provides faster convergence when getting the accurate solution on a wide range of dimensional space and is also able to solve the optimization issues with multimodal benchmark data. The evaluation of the suggested algorithm is done with various E-Commerce data and secures high accuracy on sentimental analysis and prediction while comparing with the traditional models.

An Ensembled Framework for Human Breast Cancer Survivability Prediction Using Deep Learning.

Breast cancer is categorized as an aggressive disease, and it is one of the leading causes of death. Accurate survival predictions for both long-term and short-term survivors, when delivered on time, can help physicians make effective treatment decisions for their patients. Therefore, there is a dire need to design an efficient and rapid computational model for breast cancer prognosis. In this study, we propose an ensemble model for breast cancer survivability prediction (EBCSP) that utilizes multi-modal data and stacks the output of multiple neural networks. Specifically, we design a convolutional neural network (CNN) for clinical modalities, a deep neural network (DNN) for copy number variations (CNV), and a long short-term memory (LSTM) architecture for gene expression modalities to effectively handle multi-dimensional data. The independent models' results are then used for binary classification (long term > 5 years and short term < 5 years) based on survivability using the random forest method. The EBCSP model's successful application outperforms models that utilize a single data modality for prediction and existing benchmarks.

Computational model for the semi‐analytical assessment of the free‐edge effect in composite laminated shells

AbstractThe development of efficient computational models for the accurate prediction of the state variables in general composite laminated shells undergoing uniform edge loadings is a major challenge, especially when stress concentration phenomena such as the free‐edge effect have to be considered. This paper addresses this issue by introducing a higher‐order semi‐analytical approach for the assessment of the three‐dimensional stress field in circular cylindrical composite shells with arbitrary layups subjected to a uniform bending moment. The presented semi‐analytical approach combines a closed‐form analytical plane‐strain solution with a higher‐order layerwise approach, and the governing equations are derived by virtue of the principle of minimum elastic potential. The resulting system of coupled ordinary differential equations is then solved by means of the state‐space approach and the free constants are determined by evaluating the boundary conditions at the traction‐free edges. A comparison of the numerical results of the presented semi‐analytical method with finite element simulations for various composite laminates indicates that the developed method works with high accuracy, although being extremely efficient in terms of computational resources.

Study on temperature field and influencing factors of the high geothermal tunnel with extra-long one-end construction ventilation

The increasing number of deep-buried high geothermal tunnels with extra-long one-end ventilation distance makes it critical to develop an accurate and efficient computational model to predict the ambient temperature of such tunnels. A numerical model capable of quickly calculating the longitudinal temperature field of the tunnel construction environment has been established by combining the finite difference method and finite element software. The model, which accounts for the initial ground temperature distribution as well as the heat exchanges at the working face, agrees well with the field monitoring results of the Dechang Tunnel. The research results show that the air temperature during tunnel construction varies as follows: stable heating section in the duct → rapid heating section near the working face → slow heating section of the return air → temperature drop section of the return air. The temperature of the characteristic sections (duct outlet, working face, and tunnel maximum temperature section) all show a quadratic function upward with construction length (L). These also show an exponential trend with ventilation volume (v), insulated duct thickness (t1), and tunnel insulation layer thickness (t2). To achieve a temperature change slope greater than 0.5 in the working face of the Dechang tunnel, the recommended criteria include a ventilation volume below 110 m3/s, an insulated duct thickness below 3 cm, and a tunnel insulation layer thickness below 7 cm. The research results can provide a basis for ventilation and heat insulation design of the extra-long high geothermal tunnel.

Term dependency extraction using rule-based Bayesian Network for medical image retrieval

Text-Based Medical Image Retrieval (TBMIR) has been known to be successful in retrieving medical images with textual descriptions. Usually, these descriptions are very brief and cannot express the whole visual content of the image in words, hence negatively affect the retrieval performance. One of the solutions offered in the literature is to form a Bayesian Network thesaurus taking advantage of some medical terms extracted from the image datasets. Despite the interestingness of this solution, it is not efficient as it is highly related to the co-occurrence measure, the layer arrangement and the arc directions. A significant drawback of the co-occurrence measure is the generation of a lot of uninteresting co-occurring terms. Several studies applied the association rules mining and its measures to discover the correlation between the terms. In this paper, we propose a new efficient association Rule Based Bayesian Network (R2BN) model for TBMIR using updated medically-dependent features (MDF) based on Unified Medical Language System (UMLS). The MDF are a set of medical terms that refers to the imaging modalities, the image color, the searched object dimension, etc. The proposed model presents the association rules mined from MDF in the form of Bayesian Network model. Then, it exploits the association rule measures (support, confidence, and lift) to prune the Bayesian Network model for efficient computation. The proposed R2BN model is combined with a literature probabilistic model to predict the relevance of an image to a given query. Experiments are carried out with ImageCLEF medical retrieval task collections from 2009 to 2013. Results show that our proposed model enhances significantly the image retrieval accuracy compared to the state-of-the-art retrieval models.

Generalized metabolic flux analysis framework provides mechanism-based predictions of ophthalmic complications in type 2 diabetes patients

Chronic metabolic diseases arise from changes in metabolic fluxes through biomolecular pathways and gene networks accumulated over the lifetime of an individual. While clinical and biochemical profiles present just real-time snapshots of the patients’ health, efficient computation models of the pathological disturbance of biomolecular processes are required to achieve individualized mechanistic insights into disease progression. Here, we describe the Generalized metabolic flux analysis (GMFA) for addressing this gap. Suitably grouping individual metabolites/fluxes into pools simplifies the analysis of the resulting more coarse-grain network. We also map non-metabolic clinical modalities onto the network with additional edges. Instead of using the time coordinate, the system status (metabolite concentrations and fluxes) is quantified as function of a generalized extent variable (a coordinate in the space of generalized metabolites) that represents the system’s coordinate along its evolution path and evaluates the degree of change between any two states on that path. We applied GMFA to analyze Type 2 Diabetes Mellitus (T2DM) patients from two cohorts: EVAS (289 patients from Singapore) and NHANES (517) from the USA. Personalized systems biology models (digital twins) were constructed. We deduced disease dynamics from the individually parameterized metabolic network and predicted the evolution path of the metabolic health state. For each patient, we obtained an individual description of disease dynamics and predict an evolution path of the metabolic health state. Our predictive models achieve an ROC-AUC in the range 0.79–0.95 (sensitivity 80–92%, specificity 62–94%) in identifying phenotypes at the baseline and predicting future development of diabetic retinopathy and cataract progression among T2DM patients within 3 years from the baseline. The GMFA method is a step towards realizing the ultimate goal to develop practical predictive computational models for diagnostics based on systems biology. This tool has potential use in chronic disease management in medical practice.

Modeling of heating and cooling behaviors of laminated glass facades exposed to fire with three-dimensional flexibilities

To develop a precise and efficient computer model for predicting the heating and cooling behaviors of laminated glass facades exposed to fire, there is an urgent need to reduce the huge computational requirements associated with simulating heat transfer in layered structures that feature a down-flowing water film. We overcome this challenge by proposing an efficient three-dimensional finite difference method (3DFDM) which has high numerical stability when solving heat transfer equations with water film and air convection. To capture the moving particles of the water film, we developed a unique computational algorithm for particle labeling, which has two significant advantages: (1) it eliminates the time-consuming process of searching for neighboring particles in conventional meshfree methods, and (2) it ensures that every main particle interacts only with limited neighboring particles without utilizing any weights, thus significantly reducing the computational effort. We validated our 3DFDM through experiments in heating and cooling scenarios and compared its thermal results with those obtained from commercial packages to demonstrate its high efficiency and accuracy. Furthermore, we examined the feasibility of our model in evaluating the effects of thickness of the interlayer and water film release time on the cooling behavior of laminated glass during a fire.

Analytical disk–cylinder interaction potential laws for the computational modeling of adhesive, deformable (nano)fibers

The analysis of complex fibrous systems or materials on the micro- and nanoscale, which have a high practical relevance for many technical or biological systems, requires accurate analytical descriptions of the adhesive and repulsive forces acting on the fiber surfaces. While such analytical expressions are generally needed both for theoretical studies and for computer-based simulations, the latter motivates us here to derive disk–cylinder interaction potential laws that are valid for arbitrary mutual orientations in the decisive regime of small surface separations. The chosen type of fundamental point-pair interaction follows the simple Lennard-Jones model with inverse power laws for both the adhesive van der Waals part and the steric, repulsive part. We present three different solutions, ranging from highest accuracy to the best trade-off between simplicity of the expression and sufficient accuracy for our intended use. The validity of simplifying approximations and the accuracy of the derived potential laws is thoroughly analyzed, using both numerical and analytical reference solutions for specific interaction cases. Most importantly, the correct asymptotic scaling behavior in the decisive regime of small separations is achieved, and also the theoretically predicted (1/sinα)-angle dependence (for non-parallel cylinders) is obtained by the proposed analytical solutions. As we show in the outlook to our current research, the derived analytical disk–cylinder interaction potential laws may be used to formulate highly efficient computational models for the interaction of arbitrarily curved fibers, such that the disk represents the cross-section of the first and the cylinder a local approximation to the shape of the second fiber.

A fully automatic model for premature ventricular heartbeat arrhythmia classification using the Internet of Medical Things

Cardiac arrhythmias are one of the leading causes of increased mortality worldwide and place a heavy burden on the medical environment. Premature ventricular contraction is the disturbance in electrical activity which is the most dangerous arrhythmia. Frequent occurrence of this type of arrhythmia in a regular heartbeat can lead to sudden cardiac death. Over the last decades, contemporary health-related device usage has increased the demand for efficient computational models for real-time analysis of cardiac arrhythmia. Despite notable experiments that have been done in the past decades, due to the intricate nature of PVC arrhythmia, success stories are still unsatisfying. There are numerous morphological and temporal variations present in ECG signals due to the inter-patient variability issue; extracting important characteristics of ECG signals is the most challenging task. As a result, there is a need to investigate the exact features of PVC arrhythmia, which assist in avoiding biased diagnosis. Precisely predicting it is a difficult task due to the negative polarity of PVC arrhythmia, the irregular mechanic of the ECG cycle, and anomalies between the normal cardiac rhythm. Furthermore, most of the studies in the literature followed the public benchmark dataset for the PVC arrhythmia classification, which is already pre-processed dataset. This study opens the door for a new direction of research using our unique, fully automatic model for PVC arrhythmia classification (FAPAC). This study designed an ECG monitoring module using the IoMT devices to obtain the real-time dataset for experiments and extract the relevant features from ECG signals. To classify the ECG beats, the fastest extended version of the recurrent neural network (RNN) model cyclic echo state networks to predict PVC arrhythmia. Our proposed FAPAC model successfully achieved 99.97% of accuracy, 99.99 % sensitivity,99.99% specificity, and 99.98% positive predictivity using the MIT-BIH-arrhythmia dataset, which is relatively higher than compared studies.

Petroleum Price Prediction with CNN-LSTM and CNN-GRU Using Skip-Connection

Crude oil plays an important role in the global economy, as it contributes one-third of the energy consumption worldwide. However, despite its importance in policymaking and economic development, forecasting its price is still challenging due to its complexity and irregular price trends. Although a significant amount of research has been conducted to improve forecasting using external factors as well as machine-learning and deep-learning models, only a few studies have used hybrid models to improve prediction accuracy. In this study, we propose a novel hybrid model that captures the finer details and interconnections between multivariate factors to improve the accuracy of petroleum oil price prediction. Our proposed hybrid model integrates a convolutional neural network and a recurrent neural network with skip connections and is trained using petroleum oil prices and external data directly accessible from the official website of South Korea’s national oil corporation and the official Yahoo Finance site. We compare the performance of our univariate and multivariate models in terms of the Pearson correlation, mean absolute error, mean squared error, root mean squared error, and R squared (R2) evaluation metrics. Our proposed models exhibited significantly better performance than the existing models based on long short-term memory and gated recurrent units, showing correlations of 0.985 and 0.988, respectively, for 10-day price predictions and obtaining better results for longer prediction periods when compared with other deep-learning models. We validated that our proposed model with skip connections outperforms the benchmark models and showed that the convolutional neural network using gated recurrent units with skip connections is superior to the compared models. The findings suggest that, to some extent, relying on a single source of data is ineffective in predicting long-term changes in oil prices, and thus, to develop a better prediction model based on time-series based data, it is necessary to take a multivariate approach and develop an efficient computational model with skip connections.

Multihyperuniform long-range order in medium-entropy alloys

Medium-entropy alloys (MEAs) and high-entropy alloys (HEAs) are alloys composed of mixtures of equal or relatively large proportions of three or more elements, and are distinctly different from traditional metallic alloys which contain one or two major components with smaller amounts of other elements. Recently MEAs and HEAs have received extensive research attention because they are found to possess superior mechanical performance, unusual thermal and electronic transport properties, and are resistant to corrosion; however, a fundamental understanding of the atomic structures of these alloys is still largely lacking. In this work, we develop a multihyperuniform (MH) model for disordered medium-entropy alloys (MEAs), in which the normalized infinite-wavelength composition fluctuations for all atomic species are completely suppressed, i.e., a model with hidden long-range order. Using SiGeSn alloy as a representative example, we show that this new type of highly efficient generic computational model results in stable lower-energy states compared to prevailing alloy models with (quasi)random structures, and captures experimentally observed atomic short-range orders in MEAs that are missing in (quasi)random alloy models. The MH model approximately realizes the Vegard’s law, which offers a rule-of-mixture type predictions of the lattice constants and electronic band gap, and thus can be considered as an ideal mixing state. The multihyperuniformity also directly gives rise to enhanced electronic band gaps and superior thermal transport properties at low temperatures compared to (quasi)random structures, which open up novel potential applications in optoelectronics and thermoelectrics. Our MH model can be readily applied and generalized to other medium- and high-entropy alloys (HEAs) with atomic short-range orders, and may account for some of the previously observed major discrepancies between theory and experiment.

Design of ANN Based Non-Linear Network Using Interconnection of Parallel Processor

Suspicious mass traffic constantly evolves, making network behaviour tracing and structure more complex. Neural networks yield promising results by considering a sufficient number of processing elements with strong interconnections between them. They offer efficient computational Hopfield neural networks models and optimization constraints used by undergoing a good amount of parallelism to yield optimal results. Artificial neural network (ANN) offers optimal solutions in classifying and clustering the various reels of data, and the results obtained purely depend on identifying a problem. In this research work, the design of optimized applications is presented in an organized manner. In addition, this research work examines theoretical approaches to achieving optimized results using ANN. It mainly focuses on designing rules. The optimizing design approach of neural networks analyzes the internal process of the neural networks. Practices in developing the network are based on the interconnections among the hidden nodes and their learning parameters. The methodology is proven best for nonlinear resource allocation problems with a suitable design and complex issues. The ANN proposed here considers more or less 46k nodes hidden inside 49 million connections employed on full-fledged parallel processors. The proposed ANN offered optimal results in real-world application problems, and the results were obtained using MATLAB.

Convolution Neural Networks Using Deep Matrix Factorization for Predicting Circrna-Disease Association.

CircRNAs have a stable structure, which gives them a higher tolerance to nucleases. Therefore, the properties of circular RNAs are beneficial in disease diagnosis. However, there are few known associations between circRNAs and disease. Biological experiments identify new associations is time-consuming and high-cost. As a result, there is a need of building efficient and achievable computation models to predict potential circRNA-disease associations. In this paper, we design a novel convolution neural networks framework(DMFCNNCD) to learn features from deep matrix factorization to predict circRNA-disease associations. Firstly, we decompose the circRNA-disease association matrix to obtain the original features of the disease and circRNA, and use the mapping module to extract potential nonlinear features. Then, we integrate it with the similarity information to form a training set. Finally, we apply convolution neural networks to predict the unknown association between circRNAs and diseases. The five-fold cross-validation on various experiments shows that our method can predict circRNA-disease association and outperforms state of the art methods.

On the transient performance of agglomerated graphene platelets-reinforced porous sandwich plates

We in this work present an efficient and straightforward computational model to discover transient responses of agglomerated graphene platelets (GPLs)–reinforced porous sandwich plates subjected to various dynamic loads. To this end, a generalized three-variable high-order shear deformation plate theory within the NURBS-based isogeometric analysis framework is exploited to approximate the displacements of the sandwich plates. Furthermore, a novel advanced sandwich plate model made of a porous core layer and two face layers reinforced with GPLs is proposed in this work. Two dispersion patterns of internal pores in terms of the thickness direction in the core layer including symmetric and uniform distributions are considered. Herein, a number of numerical investigations are carried out to study the effect of a wide range of key parameters on transient responses of advanced sandwich plates under four different dynamic loadings including step, triangular, half-sine and explosive blast loadings. For the first time, a significant influence of the GPLs agglomeration phenomenon on transient responses of the proposed sandwich plate model is thoroughly presented. The findings of the current work can be considered as benchmarks as well as bring useful guides for the structural design process with exceptional features in the future.

Report on the 2nd International Workshop on Transforms in Behavioral and Affective Computing (THECOG 2022) at CIKM 2022

Human decision making is central in many functions across a broad spectrum of fields including marketing, investments and smart contracts, digital health, political campaigns, logistics, and strategic management to name only a few. Computational behavioral science, the focus of the second consecutive iteration of the international workshop on transforms in behavioral and affective computing (THECOG) which was held in conjunction with CIKM 2022, not only studies the various psychological, cultural, and social factors contributing to decision making besides reasoning, but it also seeks to construct robust, scalable, and efficient computational models imitating or extending decision making processes. This year the keynote speech focused on affective robotics and their expected advantages in substantially improving the quality of human life. Moreover, the accepted papers had a considerable topical variety covering among others smart cities, speech emotion recognition, deepfake discovery, and how smart coupons may influence online consumer behavior. THECOG 2022 for a second continuous year was a central meeting point where new results were presented. Date: 21 October, 2022. Website: https://www.cikm2022.org/workshops.

Brain inspired face recognition: A computational framework

This paper presents a new proposal of an efficient computational model of face recognition which uses cues from the distributed face recognition mechanism of the brain, and by gathering engineering equivalent of these cues from existing literature. Three distinct and widely used features – Histogram of Oriented Gradients (HOG), Local Binary Patterns (LBP), and Principal components (PCs) extracted from target images are used in a manner which is simple, and yet effective. The HOG and LBP features further undergo Principal Component Analysis (PCA) for dimensionality reduction. Our model uses multi-layer perceptrons (MLP) to classify these three features and fuse them to form a sparsely connected model. A computational theory is first developed by using concepts from the information processing mechanism of the brain. Extensive experiments are carried out using eight publicly available face datasets to validate our proposed model’s performance in recognizing faces with extreme variation of illumination, pose angle, expression, and background. We also investigate the same mechanism because of reasons discussed later, on object recognition tasks as well. Results obtained are extremely promising when compared with other face and object recognition algorithms including CNN and deep learning-based methods. This highlights that simple computational processes, if clubbed properly, can produce competing performance with best algorithms.

XGBCDA: a multiple heterogeneous networks-based method for predicting circRNA-disease associations

BackgroundBiological experiments have demonstrated that circRNA plays an essential role in various biological processes and human diseases. However, it is time-consuming and costly to merely conduct biological experiments to detect the association between circRNA and diseases. Accordingly, developing an efficient computational model to predict circRNA-disease associations is urgent.MethodsIn this research, we propose a multiple heterogeneous networks-based method, named XGBCDA, to predict circRNA-disease associations. The method first extracts original features, namely statistical features and graph theory features, from integrated circRNA similarity network, disease similarity network and circRNA-disease association network, and then sends these original features to the XGBoost classifier for training latent features. The method utilizes the tree learned by the XGBoost model, the index of leaf that instance finally falls into, and the 1 of K coding to represent the latent features. Finally, the method combines the latent features from the XGBoost with the original features to train the final model for predicting the association between the circRNA and diseases.ResultsThe tenfold cross-validation results of the XGBCDA method illustrate that the area under the ROC curve reaches 0.9860. In addition, the method presents a striking performance in the case studies of colorectal cancer, gastric cancer and cervical cancer.ConclusionWith fabulous performance in predicting potential circRNA-disease associations, the XGBCDA method has the promising ability to assist biomedical researchers in terms of circRNA-disease association prediction.

Comparison of aerodynamic models for horizontal axis wind turbine blades accounting for curved tip shapes

AbstractCurved tip extensions are among the rotor innovation concepts that can contribute to the higher performance and lower cost of horizontal axis wind turbines. One of the key drivers to exploit their advantages is the use of accurate and efficient computational aerodynamic models during the design stage. The present work gives an overview of the performance of different state‐of‐the‐art models. The following tools were employed, in descending order of complexity: (i) a blade‐resolved Navier Stokes solver, (ii) a lifting line model, (iii) a vortex‐based method coupling a near‐wake model with a far‐wake model, and (iv) two implementations of the widely used blade element momentum method (BEM), with and without radial induction. The predictions of the codes were compared when simulating the baseline geometry of a reference wind turbine and different tip extension designs with relatively large sweep angle and/or dihedral angle. Four load cases were selected for this comparison, to cover several aspects of the aerodynamic modeling: steady power curve, pitch step, extreme operating gust impact, and standstill in deep stall. The present study highlighted the limitations of the BEM‐based formulations to capture the trends attributed to the introduction of curvature at the tip. This was true even when using the radial induction submodel. The rest of the computational methods showed relatively good agreement in most of the studied load cases. An exception to this was the standstill configuration, as the blade‐resolved Navier‐Stokes solver was the only code able to capture the highly unsteady effects of deep stall.

RPNet: Rotational pooling net for efficient Micro Aerial Vehicle trail navigation

Smart agile Micro Aerial Vehicles are becoming ubiquitous in industrial applications; however, their computational efficacy poses a constraint limiting their deployment. The challenge mentioned above is mitigated in the context of a Micro Aerial Vehicle vision-based navigation task. RPNet, a computational-efficient model, is proposed as a Micro Aerial Vehicle navigational controller. RPNet comprises a sequential arrangement of generic imaginary and real Gabor filters for computation reduction. Further, a novel rotational pooling mechanism that induces online feature descriptors augmentation is proposed and plugged after each convolutional block in RPNet for robustness and increase in performance. RPNet is initially trained on synthetic data for domain knowledge and further trained and tested in a real-world setting using a Micro Aerial Vehicle. Extensive experimental verification of RPNet based on four evaluation metrics shows satisfactory performance compared to the reference trajectories. Further, via comparisons, RPNet attains better error distribution of about ±5 m, and computational conservation of around 9% than the first runner-up comparator among the state-of-the-art models in the vision-based Micro Aerial Vehicle navigation task.