Number Of Training Instances Research Articles

Improved minority attack detection in Intrusion Detection System using efficient feature selection algorithms

AbstractMachine Learning and Data Mining algorithms are used extensively to enhance the performance of Intrusion Detection Systems. The number of training instances and the dimensionality of data are crucial factors affecting the performance of the model built during the training of any supervised learning algorithms. A sufficient proportion of instances having relevant features from all classes of attacks and normal traffic are considered most desirable while building the classification model that classifies the network traffic into attack and normal. This paper proposes a methodology to improve the accuracy of the model by giving importance to the relevant features that can contribute to model building. The feature selection using correlation‐based and information gain‐based techniques during training and testing contributes much to the detection of stealthier attacks and minority attacks. Then the features of the less detected attacks are identified as the second phase of the filter that is used to improve the performance. The relevant features of stealthy attacks are identified based on the correlation of corresponding features of the attack and normal data as the attacks are made stealthy mostly by making it resemble the normal traffic. Finally, the attacks that are rarely found in the training data are oversampled to improve their detection. CICIDS 2017 data set is employed as it comprises stealthier attacks generated using modern tools. NSL KDD data set is also used for evaluation to compare the proposed work with existing literature as it is used in most of the available literature. The results show superior performance with an accuracy of 99.8%, false positive rate of 0.2%, and a detection rate and 99.8%.

Classification And Errors Analysis of COVID-19 Using Supervised Machine Learning Algorithm and Visualisation

During pandemic many people died as a result of the covid-19 sickness, which appeared in 2019 andspreadovertheworld.TheobjectiveofresearchworkistowardstheoccurrenceofCOVIDtoimproveclassificationaccuracy and threshold curve predictions on real-life dataset for Receiver Operator Characteristics(ROC)value.This paper goals the real-life COVID patients from the five countries to test the experiment. The proposedmethodology involves of two steps; used Weka for calculating the accuracy by applying Decision Table machinelearning classifier and compare the results of all the five countries, secondly, the improvement in ROC value interms of initial care predictions by area under ROC analysis. For our COVID dataset has 209instances and 16attributes, Weka has performed on the number of training instances are 184,number of Rules applied is 20,searchdirection has been applied in forward direction, total number of subsets evaluated is 96,merit of best subset foundis82.609andtimetakentobuildmodelis0.06seconds.Oneadvantageofoursuggestedmodelisthatitkeepstheoriginaldataintact,ensuringexperimentquality.Afurtheradvantageisthatthemodelcanbeused withadditionaldatasetstoproducethehighest accuracyandROC analysisoutcomes

A full data augmentation pipeline for small object detection based on generative adversarial networks

Object detection accuracy on small objects, i.e., objects under 32 × 32 pixels, lags behind that of large ones. To address this issue, innovative architectures have been designed and new datasets have been released. Still, the number of small objects in many datasets does not suffice for training. The advent of the generative adversarial networks (GANs) opens up a new data augmentation possibility for training architectures without the costly task of annotating huge datasets for small objects. In this paper, we propose a full pipeline for data augmentation for small object detection which combines a GAN-based object generator with techniques of object segmentation, image inpainting, and image blending to achieve high-quality synthetic data. The main component of our pipeline is DS-GAN, a novel GAN-based architecture that generates realistic small objects from larger ones. Experimental results show that our overall data augmentation method improves the performance of state-of-the-art models up to 11.9% APs@.5 on UAVDT and by 4.7% APs@.5 on iSAID, both for the small objects subset and for a scenario where the number of training instances is limited.

Automatic document classification: the role of interclass similarity

The continuous increase of information in digital format requires new methods and techniques to access, collect and organize these volumes of textual information. One of the most widely used techniques to organize information is the automatic classification of documents. Automatic text classification systems have a low efficiency when the classes are very similar, i.e. there is overlap between them, and in this case it is very important to be able to identify those attributes that allow us to separate one class from another. In this paper we present the relationship between overlap between classes and classification accuracy. A public corpus with four classes is used for the evaluation, and each class is further separated by positives and negatives. The results obtained from four subsets with different number of training instances are presented, for each case the similarity plots, the accuracy value and the confusion matrices obtained are presented. The results obtained are very illustrative and show that the higher the similarity between classes, the lower the classification accuracy.

Convolutional neural network based prediction of effective diffusivity from microscope images

This study presents a method to predict the effective diffusivity of porous media from a limited set of scanning electron microscope images using deep learning. The electrodeposited nickel oxide film over carbon nanostructure, meant to provide electrochemical capacitance based on diffusion of ions through the film's pore space, was observed under the scanning electron microscope. Gray scale SEM images were converted to binary images, and the effective diffusivity was found using lattice Boltzmann simulations. A convolutional neural network (CNN) model comprising two sets of convolution and pooling layers and a fully connected layer was trained with lattice Boltzmann method data, where the choices of kernel size and stride were made, keeping the homogeneity of the image in perspective. An initial attempt to train a CNN with 900 training instances predicted effective diffusivity with a relative error of 13.33%, 43.49%, and 36.41% for the training, validation, and test set, respectively. However, it was over-predicting the diffusivities for images with very low connectivity. Isolated pores within the images were identified and were eliminated to solve this issue before using the images to train another CNN. This second network worked well for images with low connectivity and brought down the mean relative error from 36.41% to 25.43% for the test set. The error was found to vary across the images due to highly non-even representation of different image types, increasing by complex pore connectivity at lower porosity for which the number of training instances is the least, and overfitting of networks on medium porosity images that are present in greater numbers. The SEM images were further classified based on porosity as well as pore connectivity. The training was performed with one category of images, and the testing was conducted with images of another category at the other end of the spectrum to evaluate CNN efficacy upon encountering a new category of images. The results show the ability of the network to extend the prediction to an unknown electrode morphology, which is critical for the analysis of SEM images.

Multi-Source Domain Adaptation Techniques for Mitigating Batch Effects: A Comparative Study

The past decade has seen an increasing number of applications of deep learning (DL) techniques to biomedical fields, especially in neuroimaging-based analysis. Such DL-based methods are generally data-intensive and require a large number of training instances, which might be infeasible to acquire from a single acquisition site, especially for data, such as fMRI scans, due to the time and costs that they demand. We can attempt to address this issue by combining fMRI data from various sites, thereby creating a bigger heterogeneous dataset. Unfortunately, the inherent differences in the combined data, known as batch effects, often hamper learning a model. To mitigate this issue, techniques such as multi-source domain adaptation [Multi-source Domain Adversarial Networks (MSDA)] aim at learning an effective classification function that uses (learned) domain-invariant latent features. This article analyzes and compares the performance of various popular MSDA methods [MDAN, Domain AggRegation Networks (DARN), Multi-Domain Matching Networks (MDMN), and Moment Matching for MSDA (M3SDA)] at predicting different labels (illness, age, and sex) of images from two public rs-fMRI datasets: ABIDE 1and ADHD-200. It also evaluates the impact of various conditions such as class imbalance, the number of sites along with a comparison of the degree of adaptation of each of the methods, thereby presenting the effectiveness of MSDA models in neuroimaging-based applications.

Prompt-Based Prototypical Framework for Continual Relation Extraction

Continual relation extraction (CRE) is an important task of continual learning, which aims to learn incessantly emerging new relations between entities from texts. To avoid catastrophically forgetting old relations, some existing research efforts have focused on exploring memory replayed methods by storing typical historical learned instances or embedding all observed relations as prototypes in the episodic memory and replaying them in the subsequent training process. However, they generally fail to exploit the relation knowledge contained in the pre-trained language model (PLM), which could provide enlightening information to the representations of new relations from the known ones. To this end, we investigate the CRE from a novel perspective by generating knowledge-infused relation prototypes to leverage the relational knowledge from PLM with prompt tuning. Specifically, based on the typical samples collected from the historical learned instances with K-means algorithm, we devise novel relational knowledge-infused prompts to elicit relational knowledge from PLM for generating knowledge-infused relation prototypes. Then the prototypes are used to refine the typical examples embedding and calculate the stability-plasticity balance score for adjusting the memory replayed progress. The experimental results show that our method outperforms the state-of-the-art baseline models in CRE. The further extensive analysis presents that the proposed method is robust to memory size, task order, length of the task sequence, and the number of training instances.

Large scale multi-label learning using Gaussian processes

We introduce a Gaussian process latent factor model for multi-label classification that can capture correlations among class labels by using a small set of latent Gaussian process functions. To address computational challenges, when the number of training instances is very large, we introduce several techniques based on variational sparse Gaussian process approximations and stochastic optimization. Specifically, we apply doubly stochastic variational inference that sub-samples data instances and classes which allows us to cope with Big Data. Furthermore, we show it is possible and beneficial to optimize over inducing points, using gradient-based methods, even in very high dimensional input spaces involving up to hundreds of thousands of dimensions. We demonstrate the usefulness of our approach on several real-world large-scale multi-label learning problems.

ECG beat classification using neural classifier based on deep autoencoder and decomposition techniques

This paper proposes an ECG beat classification system based on deep autoencoder as feature extractor and a system of multiple neural networks as classifier. The objectives are as follows: First is simplifying the feature extraction step by applying the deep autoencoder, which permits defining high level features without neither pre-processing stage nor expert intervention. Second is enhancing the classification performance by decomposing the original multi-class problem into simpler binary subproblems and solving them using independent classifiers. Third is overcoming the problem of imbalanced data, by applying an oversampling method after the decomposition of the original problem. This allows adding synthetic samples according the number of training instances in each subproblem. To evaluate the proposed system, we conduct experiments on MIT-BIH arrhythmia dataset and we consider the recommendations of the Association for the Advancement of Medical Instrumentation, which defines five classes of interest. Furthermore, we perform two types of tests, i.e. intra- and inter-patient, and compare the obtained results with some of the state-of-the-art methods. We show that solving each subproblem independently can enhance the accuracy, sensitivity and specificity.

Synthesis of computer simulation and machine learning for achieving the best material properties of filled rubber

Molecular dynamics (MD) simulation is used to analyze the mechanical properties of polymerized and nanoscale filled rubber. Unfortunately, the computation time for a simulation can require several months’ computing power, because the interactions of thousands of filler particles must be calculated. To alleviate this problem, we introduce a surrogate convolutional neural network model to achieve faster and more accurate predictions. The major difficulty when employing machine-learning-based surrogate models is the shortage of training data, contributing to the huge simulation costs. To derive a highly accurate surrogate model using only a small amount of training data, we increase the number of training instances by dividing the large-scale simulation results into 3D images of middle-scale filler morphologies and corresponding regional stresses. The images include fringe regions to reflect the influence of the filler constituents outside the core regions. The resultant surrogate model provides higher prediction accuracy than that trained only by images of the entire region. Afterwards, we extract the fillers that dominate the mechanical properties using the surrogate model and we confirm their validity using MD.

Associated Learning: Decomposing End-to-End Backpropagation Based on Autoencoders and Target Propagation.

Backpropagation (BP) is the cornerstone of today's deep learning algorithms, but it is inefficient partially because of backward locking, which means updating the weights of one layer locks the weight updates in the other layers. Consequently, it is challenging to apply parallel computing or a pipeline structure to update the weights in different layers simultaneously. In this letter, we introduce a novel learning structure, associated learning (AL), that modularizes the network into smaller components, each of which has a local objective. Because the objectives are mutually independent, AL can learn the parameters in different layers independently and simultaneously, so it is feasible to apply a pipeline structure to improve the training throughput. Specifically, this pipeline structure improves the complexity of the training time from , which is the time complexity when using BP and stochastic gradient descent (SGD) for training, to , where is the number of training instances and is the number of hidden layers. Surprisingly, even though most of the parameters in AL do not directly interact with the target variable, training deep models by this method yields accuracies comparable to those from models trained using typical BP methods, in which all parameters are used to predict the target variable. Consequently, because of the scalability and the predictive power demonstrated in the experiments, AL deserves further study to determine the better hyperparameter settings, such as activation function selection, learning rate scheduling, and weight initialization, to accumulate experience, as we have done over the years with the typical BP method. In addition, perhaps our design can also inspire new network designs for deep learning. Our implementation is available at https://github.com/SamYWK/Associated_Learning.

Detection of heat pumps from smart meter and open data

Heat pumps embody solutions that heat or cool buildings effectively and sustainably, with zero emissions at the place of installation. As they pose significant load on the power grid, knowledge on their existence is crucial for grid operators, e.g., to forecast load and to plan grid operation. Further details, like the thermal reservoir (ground or air source) or the age of a heat pump installation renders energy-related services possible that utility companies can offer in the future (e.g., detecting wrongly calibrated installations, household energy efficiency checks). This study investigates the prediction of heat pump installations, their thermal reservoir and age. For this, we obtained a dataset with 397 households in Switzerland, all equipped with smart meters, collected ground truth data on installed heat pumps and enriched this data with weather data and geographical information. Our investigation replicates the state of the art in the area of heat pump detection and goes beyond it, as we obtain three major findings: First, machine learning can detect the existence of heat pumps with an AUC performance metric of 0.82, their heat reservoir with an AUC of 0.86, and their age with an AUC of 0.73. Second, heat pump existence can be better detected using data during the heating period than during summer. Third the number of training samples to detect the existence of heat pumps must not be necessarily large in terms of the number of training instances and observation period.

Identifying Clinical Study Types from PubMed Metadata: The Active (Machine) Learning Approach.

We examined a process for automating the classification of articles in MEDLINE aimed at minimising manual effort without sacrificing accuracy. From 22,808 articles pertaining to 19 antidepressants, 1000 were randomly selected and manually labelled according to article type (including, randomised controlled trials, editorials, etc.). We applied a machine learning approach termed 'active learning', where the learner (machine) selects the order in which the user (human) labels examples. Via simulation, we determined the number of articles a user needed to label to produce a classifier with at least 95% recall and 90% precision in three scenarios related to evidence synthesis. We found that the active learning process reduced the number of training instances required by 70%, 19%, and 14% in the three scenarios. The results show that the active learning method may be used in some scenarios to produce accurate classifiers that meet the needs of evidence synthesis tasks and reduce manual effort.

Efficient K-Nearest Neighbor Searches for Multiple-Face Recognition in the Classroom based on Three Levels DWT-PCA

The main weakness of the k-Nearest Neighbor algorithm in face recognition is calculating the distance and sort all training data on each prediction which can be slow if there are a large number of training instances. This problem can be solved by utilizing the priority k-d tree search to speed up the process of k-NN classification. This paper proposes a method for student attendance systems in the classroom using facial recognition techniques by combining three levels of Discrete Wavelet Transforms (DWT) and Principal Component Analysis (PCA) to extract facial features followed by applying the priority of k-d tree search to speed up the process of facial classification using k-Nearest Neighbor. The proposed algorithm is tested on two datasets that are Honda/UCSD video dataset and our dataset (AtmafaceDB dataset). This research looks for the best value of k to get the right facial recognition using k-fold cross-validation. 10-fold cross-validation at level 3 DWT-PCA shows that face recognition using k-Nearest Neighbor on our dataset is 95.56% with k = 5, whereas in the Honda / UCSD dataset it is only 82% with k = 3. The proposed method gives computational recognition time on our dataset 40 milliseconds.

Automatically Evolving Rotation-invariant Texture Image Descriptors by Genetic Programming

In computer vision, training a model that performs classification effectively is highly dependent on the extracted features, and the number of training instances. Conventionally, feature detection and extraction are performed by a domain expert who, in many cases, is expensive to employ and hard to find. Therefore, image descriptors have emerged to automate these tasks. However, designing an image descriptor still requires domain-expert intervention. Moreover, the majority of machine learning algorithms require a large number of training examples to perform well. However, labeled data is not always available or easy to acquire, and dealing with a large dataset can dramatically slow down the training process. In this paper, we propose a novel genetic programming-based method that automatically synthesises a descriptor using only two training instances per class. The proposed method combines arithmetic operators to evolve a model that takes an image and generates a feature vector. The performance of the proposed method is assessed using six datasets for texture classification with different degrees of rotation and is compared with seven domain-expert designed descriptors. The results show that the proposed method is robust to rotation and has significantly outperformed, or achieved a comparable performance to, the baseline methods.

Expectation propagation in linear regression models with spike-and-slab priors

An expectation propagation (EP) algorithm is proposed for approximate inference in linear regression models with spike-and-slab priors. This EP method is applied to regression tasks in which the number of training instances is small and the number of dimensions of the feature space is large. The problems analyzed include the reconstruction of genetic networks, the recovery of sparse signals, the prediction of user sentiment from customer-written reviews and the analysis of biscuit dough constituents from NIR spectra. The proposed EP method outperforms in most of these tasks another EP method that ignores correlations in the posterior and a variational Bayes technique for approximate inference. Additionally, the solutions generated by EP are very close to those given by Gibbs sampling, which can be taken as the gold standard but can be much more computationally expensive. In the tasks analyzed, spike-and-slab priors generally outperform other sparsifying priors, such as Laplace, Student's $$t$$t and horseshoe priors. The key to the improved predictions with respect to Laplace and Student's $$t$$t priors is the superior selective shrinkage capacity of the spike-and-slab prior distribution.

Towards More Accurate Clutering Method by Using Dynamic Time Warping

An intrinsic problem of classifiers based on machine learning (ML) methods is that their learning time grows as the size and complexity of the training dataset increases. For this reason, it is important to have efficient computational methods and algorithms that can be applied on large datasets, such that it is still possible to complete the machine learning tasks in reasonable time. In this context, we present in this paper a more accurate simple process to speed up ML methods. An unsupervised clustering algorithm is combined with Expectation, Maximization (EM) algorithm to develop an efficient Hidden Markov Model (HMM) training. The idea of the proposed process consists of two steps. In the first step, training instances with similar inputs are clustered and a weight factor which represents the frequency of these instances is assigned to each representative cluster. Dynamic Time Warping technique is used as a dissimilarity function to cluster similar examples. In the second step, all formulas in the classical HMM training algorithm (EM) associated with the number of training instances are modified to include the weight factor in appropriate terms. This process significantly accelerates HMM training while maintaining the same initial, transition and emission probabilities matrixes as those obtained with the classical HMM training algorithm. Accordingly, the classification accuracy is preserved. Depending on the size of the training set, speedups of up to 2200 times is possible when the size is about 100.000 instances. The proposed approach is not limited to training HMMs, but it can be employed for a large variety of MLs methods.

SET EXPANSION OF CONTEXTUAL SEMANTIC RELATIONS: AN ALTERNATIVE TO FULL CORPUS ANNOTATION FOR SUPERVISED CLASSIFICATION

Recent approaches for classification of semantic relations are based on supervised learning using large training datasets. Due to the high cost of annotating such data and to the class imbalance problem, alternatives for minimizing the effort of full corpus annotation are required. In set expansion, one of such alternatives, given a small initial training set, new relevant instances are acquired from a large corpus. However, when dealing with contextual semantic relations, which are relations that are highly dependent on the context within the sentence, set expansion is not trivial, since instances are not directly queryable and filtering requires classification under a very restricted number of training instances. This work thus proposes a bootstrapped set expansion method for contextual semantic relations. It performs a best effort extraction using the Web, and a two-stage filtering of candidate instances, the first based on syntactic patterns and the second using a feature distance-based classifier designed for the low frequency setting. The relevance of the output is measured experimentally by using the expanded set as the training data of the supervised classification task, observing an incremental improvement in performance after each bootstrapping iteration when compared to values using the unexpanded training data.

Innocent Until Proven Guilty: Reducing Robot Shaping From Polynomial to Linear Time

In evolutionary algorithms, much time is spent evaluating inferior phenotypes that produce no offspring. A common heuristic to address this inefficiency is to stop evaluations early if they hold little promise of attaining high fitness. However, the form of this heuristic is typically dependent on the fitness function used, and there is a danger of prematurely stopping evaluation of a phenotype that may have recovered in the remainder of the evaluation period. Here a stopping method is introduced that gradually reduces fitness over the phenotype's evaluation, rather than accumulating fitness. This method is independent of the fitness function used, only stops those phenotypes that are guaranteed to become inferior to the current offspring-producing phenotypes, and realizes significant time savings across several evolutionary robotics tasks. It was found that for many tasks, time complexity was reduced from polynomial to sublinear time, and time savings increased with the number of training instances used to evaluate a phenotype as well as with task difficulty.

An Approach to Multi-Pitch Estimation Using a Support Vector Machine

In this paper, an approach to multi-pitch estimation using a support vector machine (SVM) is proposed. For automatic music transcription, an SVM is used often for identifying the sounds of musical instruments but rarely for estimation of the pitch – the fundamental frequency. Our approach is based on the SVM characteristic of finding the ideal hypersurface that separates the input vectors into two classes. The proposed system achieves multi-pitch estimation by effectively using the margins between the hypersurfaces and the input vectors. In addition, our system does not require chord sounds for the SVM training instances. Experiments using sound data generated by artificially compounding real audio signals show that the proposed method drastically reduces the number of training instances.