Downloads & Free Reading Options - Results

Reproduction report and material.

Perma.cc archive of https://towardsdatascience.com/types-of-machine-learning-algorithms-you-should-know-953a08248861 created on 2020-01-21 01:31:24+00:00.

The problem of distributed learning and channel access is considered in a cognitive network with multiple secondary users. The availability statistics of the channels are initially unknown to the secondary users and are estimated using sensing decisions. There is no explicit information exchange or prior agreement among the secondary users. We propose policies for distributed learning and access which achieve order-optimal cognitive system throughput (number of successful secondary transmissions) under self play, i.e., when implemented at all the secondary users. Equivalently, our policies minimize the regret in distributed learning and access. We first consider the scenario when the number of secondary users is known to the policy and prove that the total regret is logarithmic in the number of transmission slots. Our distributed learning and access policy achieves order-optimal regret by comparing to an asymptotic lower bound for regret under any uniformly-good learning and access policy. We then consider the case when the number of secondary users is fixed but unknown, and is estimated through feedback. We propose a policy in this scenario whose asymptotic sum regret which grows slightly faster than logarithmic in the number of transmission slots.

The European literature on algorithms for learning and instruction is reviewed in this document. The word "algorithm" is defined, the relationship between the European literature and current trends in research on learning and instruction in the United States is described, the important practical uses of algorithms are discussed, and potential high-yield research activities related to the use of algorithms are suggested. (DT)

turing test polikar mind/umer idiot

We have successfully implemented a two-stage event correlation model for intrusion detection system (IDS) alerts. The model is designed to automate alert and incidents management and reduce the workload on an IDS analyst. We achieve this correlation by clustering similar alerts together, thus allowing the analyst to only look at a few clusters instead of hundreds or thousands of alerts. The first stage of this model uses an artificial neural network (ANN)-based autoassociator. The autoassociator is trained to reproduce each alert at its output, and it uses the error metric between its input and output to cluster similar alerts together. The accuracy of the system is improved by adding another machine-learning stage which attempts to combine closely related clusters produced by the first stage into super-clusters. The second stage uses the Expectation Maximisation (EM) clustering algorithm. The model and performance of this model are tested with intrusion alerts generated by a Snort IDS on DARPA's 1999 IDS evaluation data as well as incidents.org alerts.

Considering the harms of air pollution on human health and the environment, it seems necessary to reduce and solve this problem based on accurate knowledge of pollutants and criteria affecting it and identifying polluted areas. Therefore, using mathematical models in the form of machine learning is an optimal and cost-efficient approach to air pollution modeling. This research is applied in terms of purpose and its method is descriptive-analytical. The novelty of this research is presenting a new combination approach to determine the effective criteria for predicting the amount of air pollution. Therefore, the purpose of this study was to evaluate and compare the capabilities of two machine learning models, namely Support Vector Machine (SVM) and Random Forest (RF) in combination with Genetic Algorithm (GA) to predict air pollution in Tehran. The data used in this research include particulate matter and gaseous pollutants in Tehran in 2020, which was obtained from Tehran Traffic Control Company. MATLAB and ArcMap software were used to analyze the data. The value of coefficient of determination (R2 ) obtained from the combined RF-GA method was 0.997, which indicates the high compatibility of this model with the data of this study. Moreover, the Root Mean Square Error (RMSE) value from the combined RF-GA method was 0.153, which indicates high accuracy of this model. Based on the data obtained from Tehran Traffic Control Company, the results of the RF method indicate the appropriateness of selecting the model to estimate the amount of air pollution in Tehran

Perma.cc archive of https://journals.vilniustech.lt/index.php/IJSPM/article/view/12742/9995 created on 2022-09-08 23:35:54.953297+00:00.

Machine learning is recognized as a tool for improving the performance of many kinds of systems, yet most machine learning systems themselves are not well equipped to improve their own learning performance. By emphasizing the role of domain knowledge, learning systems can be crafted as knowledge directed systems, and with the addition of a knowledge store for organizing and maintaining knowledge to assist learning, a learning machine learning (L-ML) algorithm is possible. The necessary components of L-ML systems are presented along with several case descriptions of existing machine learning systems that possess limited L-ML capabilities. Keywords: Algorithms, Artificial intelligence. (SDW)

Malaria continues to be a major public health challenge in sub-Saharan Africa, where accurate and timely diagnosis is often hindered by limitations in traditional diagnostic methods. This review evaluated the role of machine learning (ML) algorithms in improving malaria diagnosis in primary healthcare settings. Specifically, it explores applications of ML in microscopic image analysis, rapid diagnostic test (RDT) optimization, and predictive modeling, with a focus on their potential to enhance diagnostic accuracy and decision-making in resource-limited environments. ML techniques, such as convolutional neural networks (CNNs) for image analysis and data-driven models for optimizing RDT interpretation, have shown promise in addressing inter-observer variability and improving test sensitivity and specificity. Furthermore, predictive modeling integrating clinical, demographic, and environmental data can help prioritize malaria cases and guide healthcare providers in making accurate diagnoses. Despite these advancements, challenges such as data limitations, infrastructure gaps, and ethical considerations remain significant barriers to widespread adoption. The methodology utilized in this review involved a comprehensive synthesis of current literature, examining empirical studies on ML applications in malaria diagnosis and assessing their feasibility in primary healthcare contexts. To overcome these challenges, the article suggested policy recommendations, including investments in data infrastructure, capacity building, and public-private partnerships. Ultimately, ML offers a promising solution to enhance malaria diagnostic capabilities, contributing to better health outcomes in endemic regions. 

This article is from Frontiers in Public Health , volume 2 . Abstract Background: Active travel is an important area in physical activity research, but objective measurement of active travel is still difficult. Automated methods to measure travel behaviors will improve research in this area. In this paper, we present a supervised machine learning method for transportation mode prediction from global positioning system (GPS) and accelerometer data.Methods: We collected a dataset of about 150 h of GPS and accelerometer data from two research assistants following a protocol of prescribed trips consisting of five activities: bicycling, riding in a vehicle, walking, sitting, and standing. We extracted 49 features from 1-min windows of this data. We compared the performance of several machine learning algorithms and chose a random forest algorithm to classify the transportation mode. We used a moving average output filter to smooth the output predictions over time.Results: The random forest algorithm achieved 89.8% cross-validated accuracy on this dataset. Adding the moving average filter to smooth output predictions increased the cross-validated accuracy to 91.9%.Conclusion: Machine learning methods are a viable approach for automating measurement of active travel, particularly for measuring travel activities that traditional accelerometer data processing methods misclassify, such as bicycling and vehicle travel.

There are several confounding factors that can reduce the accuracy of gait recognition systems. These factors can reduce the distinctiveness, or alter the features used to characterise gait, they include variations in clothing, lighting, pose and environment, such as the walking surface. Full invariance to all confounding factors is challenging in the absence of high-quality labelled training data. We introduce a simulation-based methodology and a subject-specific dataset which can be used for generating synthetic video frames and sequences for data augmentation. With this methodology, we generated a multi-modal dataset. In addition, we supply simulation files that provide the ability to simultaneously sample from several confounding variables. The basis of the data is real motion capture data of subjects walking and running on a treadmill at different speeds. Results from gait recognition experiments suggest that information about the identity of subjects is retained within synthetically generated examples. The dataset and methodology allow studies into fully-invariant identity recognition spanning a far greater number of observation conditions than would otherwise be possible.

Network has brought convenience to the earth by permitting versatile transformation of information, however it conjointly exposes a high range of vulnerabilities. A Network Intrusion Detection System helps network directors and system to view network security violation in their organizations. Characteristic unknown and new attacks are one of the leading challenges in Intrusion Detection System researches. Deep learning that a subfield of machine learning cares with algorithms that are supported the structure and performance of brain known as artificial neural networks. The improvement in such learning algorithms would increase the probability of IDS and the detection rate of unknown attacks. Throughout, we have a tendency to suggest a deep learning approach to implement increased IDS and associate degree economical. Priya N | Ishita Popli "Comparative Study on Machine Learning Algorithms for Network Intrusion Detection System" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-1 , December 2020, URL: https://www.ijtsrd.com/papers/ijtsrd38175.pdf Paper URL : https://www.ijtsrd.com/computer-science/computer-network/38175/comparative-study-on-machine-learning-algorithms-for-network-intrusion-detection-system/priya-n

This is an outstanding book by Prof. David MacKay (of U. of Cambridge). It is downloadable from author's web page: http://www.inference.phy.cam.ac.uk/mackay/ . Please spread the word, and tell your profs to use this free book in their courses. This is an e-book free to read and share electronically as indicated so by the author. However, you are not allowed to print the whole book, instead you should order it from the publisher. Copyright Cambridge University Press 2003. On-screen viewing permitted. Printing not permitted. http://www.cambridge.org/0521642981 You can buy this book for 30 pounds or $50. See http://www.inference.phy.cam.ac.uk/mackay/itila/ for links. Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v 1 Introduction to Information Theory . . . . . . . . . . . . . 3 2 Probability, Entropy, and Inference . . . . . . . . . . . . . . 22 3 More about Inference . . . . . . . . . . . . . . . . . . . . . 48 I Data Compression . . . . . . . . . . . . . . . . . . . . . . 65 4 The Source Coding Theorem . . . . . . . . . . . . . . . . . 67 5 Symbol Codes . . . . . . . . . . . . . . . . . . . . . . . . . 91 6 Stream Codes . . . . . . . . . . . . . . . . . . . . . . . . . . 110 7 Codes for Integers . . . . . . . . . . . . . . . . . . . . . . . 132 II Noisy-Channel Coding . . . . . . . . . . . . . . . . . . . . 137 8 Dependent Random Variables . . . . . . . . . . . . . . . . . 138 9 Communication over a Noisy Channel . . . . . . . . . . . . 146 10 The Noisy-Channel Coding Theorem . . . . . . . . . . . . . 162 11 Error-Correcting Codes and Real Channels . . . . . . . . . 177 III Further Topics in Information Theory . . . . . . . . . . . . . 191 12 Hash Codes: Codes for Efficient Information Retrieval . . 193 13 Binary Codes . . . . . . . . . . . . . . . . . . . . . . . . . 206 14 Very Good Linear Codes Exist . . . . . . . . . . . . . . . . 229 15 Further Exercises on Information Theory . . . . . . . . . . 233 16 Message Passing . . . . . . . . . . . . . . . . . . . . . . . . 241 17 Communication over Constrained Noiseless Channels . . . 248 18 Crosswords and Codebreaking . . . . . . . . . . . . . . . . 260 19 Why have Sex? Information Acquisition and Evolution . . 269 IV Probabilities and Inference . . . . . . . . . . . . . . . . . . 281 20 An Example Inference Task: Clustering . . . . . . . . . . . 284 21 Exact Inference by Complete Enumeration . . . . . . . . . 293 22 Maximum Likelihood and Clustering . . . . . . . . . . . . . 300 23 Useful Probability Distributions . . . . . . . . . . . . . . . 311 24 Exact Marginalization . . . . . . . . . . . . . . . . . . . . . 319 25 Exact Marginalization in Trellises . . . . . . . . . . . . . . 324 26 Exact Marginalization in Graphs . . . . . . . . . . . . . . . 334 27 Laplace’s Method . . . . . . . . . . . . . . . . . . . . . . . 341 28 Model Comparison and Occam’s Razor . . . . . . . . . . . 343 29 Monte Carlo Methods . . . . . . . . . . . . . . . . . . . . . 357 30 Efficient Monte Carlo Methods . . . . . . . . . . . . . . . . 387 31 Ising Models . . . . . . . . . . . . . . . . . . . . . . . . . . 400 32 Exact Monte Carlo Sampling . . . . . . . . . . . . . . . . . 413 33 Variational Methods . . . . . . . . . . . . . . . . . . . . . . 422 34 Independent Component Analysis and Latent Variable Mod- elling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437 35 Random Inference Topics . . . . . . . . . . . . . . . . . . . 445 36 Decision Theory . . . . . . . . . . . . . . . . . . . . . . . . 451 37 Bayesian Inference and Sampling Theory . . . . . . . . . . 457 V Neural networks . . . . . . . . . . . . . . . . . . . . . . . . 467 38 Introduction to Neural Networks . . . . . . . . . . . . . . . 468 39 The Single Neuron as a Classifier . . . . . . . . . . . . . . . 471 40 Capacity of a Single Neuron . . . . . . . . . . . . . . . . . . 483 41 Learning as Inference . . . . . . . . . . . . . . . . . . . . . 492 42 Hopfield Networks . . . . . . . . . . . . . . . . . . . . . . . 505 43 Boltzmann Machines . . . . . . . . . . . . . . . . . . . . . . 522 44 Supervised Learning in Multilayer Networks . . . . . . . . . 527 45 Gaussian Processes . . . . . . . . . . . . . . . . . . . . . . 535 46 Deconvolution . . . . . . . . . . . . . . . . . . . . . . . . . 549 VI Sparse Graph Codes . . . . . . . . . . . . . . . . . . . . . 555 47 Low-Density Parity-Check Codes . . . . . . . . . . . . . . 557 48 Convolutional Codes and Turbo Codes . . . . . . . . . . . . 574 49 Repeat–Accumulate Codes . . . . . . . . . . . . . . . . . . 582 50 Digital Fountain Codes . . . . . . . . . . . . . . . . . . . . 589 VII Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . 597 A Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598 B Some Physics . . . . . . . . . . . . . . . . . . . . . . . . . . 601 C Some Mathematics . . . . . . . . . . . . . . . . . . . . . . . 605 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 620

We demonstrate that it is possible to measure metallicity from the SDSS five-band photometry to better than 0.1 dex using supervised machine learning algorithms. Using spectroscopic estimates of metallicity as ground truth, we build, optimize and train several estimators to predict metallicity. We use the observed photometry, as well as derived quantities such as stellar mass and photometric redshift, as features, and we build two sample data sets at median redshifts of 0.103 and 0.218 and median r-band magnitude of 17.5 and 18.3 respectively. We find that ensemble methods, such as Random Forests of Trees and Extremely Randomized Trees, and Support Vector Machines all perform comparably well and can measure metallicity with a Root Mean Square Error (RMSE) of 0.081 and 0.090 for the two data sets when all objects are included. The fraction of outliers (objects for which |Z_true - Z_pred| > 0.2 dex) is 2.2 and 3.9%, respectively and the RMSE decreases to 0.068 and 0.069 if those objects are excluded. Because of the ability of these algorithms to capture complex relationships between data and target, our technique performs better than previously proposed methods that sought to fit metallicity using an analytic fitting formula, and has 3x more constraining power than SED fitting-based methods. Additionally, this method is extremely forgiving of contamination in the training set, and can be used with very satisfactory results for training sample sizes of just a few hundred objects. We distribute all the routines to reproduce our results and apply them to other data sets.

This study evaluated the nature-inspired optimization algorithms to improve classification involving imbalanced class problems. The particle swarm optimization (PSO) and grey wolf optimizer (GWO) were used to adaptively balance the distribution and then four supervised machine learning classifiers artificial neural network (ANN), support vector machine (SVM), extreme gradient-boosted tree (XGBoost), and random forest (RF) were applied to maximize the classification performance for electricity fraud prediction. The imbalance data was balanced using random undersampling (RUS) and two nature-inspired algorithm techniques (PSO and GWO). Results showed that for the data balanced using random undersampling, ANN (Sentest = 50.31%), and XGBoost (Sentest = 66.32%) has better sensitivity than SVM (Sentest = 23.61%), while RF exhibits overfitting (Sentrain = 100%, Sentest = 71.25%). The classification performance of RF model hybrid with PSO improved tremendously (AccTest = 96.98%, Sentest = 94.87%, Spectest = 99.16%, Pretest = 99.14%, F1 Score = 96.96%, and area under the curve (AUC) = 0.989). This was closely followed by hybrid of XGBoost with PSO. Moreover, RF and XGBoost hybrid with GWO also showed an improvement and promising results. This study has showed that nature-inspired optimization algorithms (PSO and GWO) are effective methods in addressing imbalanced dataset.

This is a companion note to our recent study of the weak convergence properties of constrained emphatic temporal-difference learning (ETD) algorithms from a theoretic perspective. It supplements the latter analysis with simulation results and illustrates the behavior of some of the ETD algorithms using three example problems.

In this thesis, we propose two new machine learning schemes, a subband-based Independent Component Analysis scheme and a hybrid Independent Component Analysis/Support Vector Machine scheme, and apply them to the problems of blind acoustic signal separation and face detection. Based on a linear model, classical Independent Component Analysis (ICA) provides a method of representing data as independent components. In contrast to Principal Component Analysis (PCA), which decorrelates the data based on its covariance matrix, ICA uses higher-order statistics of the data to minimize the dependence between the components of the system output. An important application of ICA is blind source separation. However, classical ICA algorithms do not work well for separation in the presence of noise or when performed on-line. Inspired by the psychoacoustic discovery that humans perceive and process acoustic signals in different frequency bands independently, we propose a new algorithm, subband-based ICA, that integrates ICA with time-frequency analysis to separate mixed signals. In subband-based ICA, the separations are performed in parallel in several frequency bands. Wavelet decomposition and best basis selection in wavelet/DCT packets can be incorporated into this algorithm.

Talk by Marc Wanner - Predicting Ground State Properties: Constant Sample Complexity and Deep Learning Algorithms @QTMLConference

A method is being developed that provides for an artificial-intelligence system to learn a user's preferences for sets of objects and to thereafter automatically select subsets of objects according to those preferences. The method was originally intended to enable automated selection, from among large sets of images acquired by instruments aboard spacecraft, of image subsets considered to be scientifically valuable enough to justify use of limited communication resources for transmission to Earth. The method is also applicable to other sets of objects: examples of sets of objects considered in the development of the method include food menus, radio-station music playlists, and assortments of colored blocks for creating mosaics. The method does not require the user to perform the often-difficult task of quantitatively specifying preferences; instead, the user provides examples of preferred sets of objects. This method goes beyond related prior artificial-intelligence methods for learning which individual items are preferred by the user: this method supports a concept of setbased preferences, which include not only preferences for individual items but also preferences regarding types and degrees of diversity of items in a set. Consideration of diversity in this method involves recognition that members of a set may interact with each other in the sense that when considered together, they may be regarded as being complementary, redundant, or incompatible to various degrees. The effects of such interactions are loosely summarized in the term portfolio effect. The learning method relies on a preference representation language, denoted DD-PREF, to express set-based preferences. In DD-PREF, a preference is represented by a tuple that includes quality (depth) functions to estimate how desired a specific value is, weights for each feature preference, the desired diversity of feature values, and the relative importance of diversity versus depth. The system applies statistical concepts to estimate quantitative measures of the user s preferences from training examples (preferred subsets) specified by the user. Once preferences have been learned, the system uses those preferences to select preferred subsets from new sets. The method was found to be viable when tested in computational experiments on menus, music playlists, and rover images. Contemplated future development efforts include further tests on more diverse sets and development of a sub-method for (a) estimating the parameter that represents the relative importance of diversity versus depth, and (b) incorporating background knowledge about the nature of quality functions, which are special functions that specify depth preferences for features.

Overall, life-time prevalence rates of mental health problems are around 30-50% in many countries (Andrade et al., 2000; Kessler et al., 2007), imposing a heavy burden on individuals, families, and communities, accompanied with high health costs related to screening, prevention, and treatment (GBD 2019 Mental Disorders Collaborators, 2022). Previous studies have built prediction models to be able to increase detection and prevention success, and to increase knowledge on possible risk factors of mental illness (Dwyer et al., 2018; Macalli et al., 2021; Tate et al., 2022; H. Yang et al., 2010). Mental health, however, includes both mental illness and well-being. That is, well-being is not simply the absence of mental illness (Keyes, 2002). To assure that our society remains resilient it is therefore also important to develop optimal risk prediction models for well-being, to be able to predict who will thrive and understand why this is the case (Oparina et al., 2022). This information can be valuable for well-being interventions. Previous research on mental health issues have provided us with possible risk factors, that are also relevant for well-being. First, mental health in adulthood has its developmental origins in childhood and adolescence, as indicated by associations with childhood psychopathology, making the availability of longitudinal data crucial (Lahey et al., 2014; Rutter et al., 2006). Second, mental health traits (e.g., depression, life satisfaction, positive affect) are partly driven by thousands of genetic variants with many small but relevant effects, many of which are shared across disorders (Baselmans, van de Weijer, et al., 2019; Kim et al., 2022; Meng et al., 2022; Thorp et al., 2021). Third, many environmental exposures are associated with mental health, examples including socio-economic status, childhood maltreatment, substance use, urbanicity and life events (Uher & Zwicker, 2017). Just as is seen for genetic effects, environmental effects for mental health and well-being overlap. Finally, environmental factors interact with genetic effects on mental health (Assary et al., 2018; Dunn et al., 2016; Uher & Zwicker, 2017). Together, a complex picture of mental health development emerges. Optimal prediction thus likely requires a broad inclusion of possible and protective risk factors, which may lead to the identification of the most relevant factors associated with mental health. This in turn could lead to individualized prediction models for individuals’ future mental health status (Bzdok et al., 2021). Given the multitude of factors associated with mental health, accurate prediction requires appropriate methods that can deal with high complexity. The rise of big data has led to the development of machine learning methods that enable the inclusion of large numbers of variables, while accounting for their potential interactions, consistent with the consensus that mental health results from complex interactions between developmental, social, psychological, genetic, and environmental factors. Recent developments in digitalization and record linkage have further made it increasingly possible to expand our environmental scope by including more objective environmental exposures (e.g., air pollution, green spaces) in mental health models (van de Weijer et al., 2021). Previous machine learning studies on responses to anti-depressants (Taliaz et al., 2021), rehospitalization after depressive episodes (Cearns et al., 2019), and resilience after cancer diagnoses (Kourou et al., 2021) have indeed shown that models including different data modalities outperform models including a single set of predictors. In line with the multi-factorial nature of well-being, a recent study further found that expanding the set of predictive features increased the performance of the models for well-being considerably (Oparina et al., 2022). Recent developments have thus paved the way for more accurate predictions for mental health related traits (Dwyer et al., 2018). However, many studies are conducted using clinical samples, i.e., when treatment is already sought, limiting their external validity and practical usefulness, especially for prevention. In addition, most studies focused on mental illness, rather than on mental health and well-being (Macalli et al., 2021; Tate et al., 2022). At the same time, machine learning prediction studies in population samples largely failed to take an integrative approach meaning that either cross-sectional data were used, or environmental exposures and/or genetic data were limited or missing (Dwyer et al., 2018; Macalli et al., 2021; Oparina et al., 2022; Tate et al., 2022; H. Yang et al., 2010). This may explain why predictive accuracies have not reached the standards needed for clinical use yet (Runeson et al., 2017). In the current project, we will overcome these caveats by building prediction models for well-being with extensive longitudinal data on environmental and psychosocial factors, and genetic data. More specifically, by using an extensive set of predictors and utilizing novel machine learning methods that enable the combined use of multiple prediction models (stacked ensemble model; see below), we aim to build a highly generalizable, comprehensive prediction model for well-being, which can inform future models for clinical prediction and decision-making, hereby preparing society for future mental health challenges.

We investigate the OpenMP parallelization and optimization of two novel data classification algorithms. The new algorithms are based on graph and PDE solution techniques and provide significant accuracy and performance advantages over traditional data classification algorithms in serial mode. The methods leverage the Nystrom extension to calculate eigenvalue/eigenvectors of the graph Laplacian and this is a self-contained module that can be used in conjunction with other graph-Laplacian based methods such as spectral clustering. We use performance tools to collect the hotspots and memory access of the serial codes and use OpenMP as the parallelization language to parallelize the most time consuming parts. Where possible, we also use library routines. We then optimize the OpenMP implementations and detail the performance on traditional supercomputer nodes (in our case a Cray XC30), and predict behavior on emerging testbed systems based on Intel's Knights Corner and Landing processors. We show both performance improvement and strong scaling behavior. A large number of optimization techniques and analyses are necessary before the algorithm reaches almost ideal scaling.

This paper gives specific divergence examples of value-iteration for several major Reinforcement Learning and Adaptive Dynamic Programming algorithms, when using a function approximator for the value function. These divergence examples differ from previous divergence examples in the literature, in that they are applicable for a greedy policy, i.e. in a "value iteration" scenario. Perhaps surprisingly, with a greedy policy, it is also possible to get divergence for the algorithms TD(1) and Sarsa(1). In addition to these divergences, we also achieve divergence for the Adaptive Dynamic Programming algorithms HDP, DHP and GDHP.

In this paper, the researchers studied the effects of different activation functions in hidden layers and how they impact the overfitting or underfitting of the model in the multiclass prediction of chronic diseases. This paper also evaluated the effects of varying the number of layers, and hyperparameters and its impact on the accuracy of the model and its generalization capabilities. It was found that exponential linear unit (ELU) does not have a significant advantage over rectified linear unit (ReLU) when used as an activation function in the hidden layer. Additionally, the performance of softmax function, when used in the output layer, is the same as a classic sigmoid output activation function. In terms of the ability of the model to generalize, the researchers achieved a classification accuracy of 100% when the trained model was used to predict unseen data. Through this research, the researchers should be able to assist medical professionals and practitioners in Oman in the validation and diagnosis of chronic diseases in clinics and hospitals.

This one-year effort will focus on the transition of FERI's machine learning algorithms for HyperSpectral Imagery (HSI) in the VSW into a distributable code set. This will provide a stable code platform for the application and transition of machine learning-based hyperspectral classification techniques into 6.3/6.4 programs. (This work was funded mid-year 2008.) Our objective is to focus on three areas of application research and transitions. First, we will transition our machine learning-based algorithms and computer code for the determination of bathymetry, bottom type, and water column Inherent Optical Properties from HyperSpectral Imagery (HSI) into a deliverable Message Passing Interface (MPI) program that may be easily used by other research and military operators. Second, we will use this program to determine the impacts of the granularity of the classification database on the inversion bathymetry, bottom type, and IOPs. Third, we will move beyond the use of single pixel HSI inversion to the use of spatial context-filtering to remove pixel-topixel noise inherent in the HSI data.

The research performed under this grant investigated three primary areas. First, collective excitation and distributed winner-take-all dynamics were investigated in laterally connected networks to further characterize the properties of biological self-organization. Self-organization was further investigated as part of the k-means clustering algorithm, where the trade-off between learning of new exemplars and global efficiency was optimized. Finally, a cross-validation technique, referred to as Generalized Prediction Error (GPE), was investigated as a means of predicting generalization error after training.

The fundamental problem of multiple secondary users contending for opportunistic spectrum access over multiple channels in cognitive radio networks has been formulated recently as a decentralized multi-armed bandit (D-MAB) problem. In a D-MAB problem there are $M$ users and $N$ arms (channels) that each offer i.i.d. stochastic rewards with unknown means so long as they are accessed without collision. The goal is to design a decentralized online learning policy that incurs minimal regret, defined as the difference between the total expected rewards accumulated by a model-aware genie, and that obtained by all users applying the policy. We make two contributions in this paper. First, we consider the setting where the users have a prioritized ranking, such that it is desired for the $K$-th-ranked user to learn to access the arm offering the $K$-th highest mean reward. For this problem, we present the first distributed policy that yields regret that is uniformly logarithmic over time without requiring any prior assumption about the mean rewards. Second, we consider the case when a fair access policy is required, i.e., it is desired for all users to experience the same mean reward. For this problem, we present a distributed policy that yields order-optimal regret scaling with respect to the number of users and arms, better than previously proposed policies in the literature. Both of our distributed policies make use of an innovative modification of the well known UCB1 policy for the classic multi-armed bandit problem that allows a single user to learn how to play the arm that yields the $K$-th largest mean reward.

The fundamental problem of multiple secondary users contending for opportunistic spectrum access over multiple channels in cognitive radio networks has been formulated recently as a decentralized multi-armed bandit (D-MAB) problem. In a D-MAB problem there are $M$ users and $N$ arms (channels) that each offer i.i.d. stochastic rewards with unknown means so long as they are accessed without collision. The goal is to design a decentralized online learning policy that incurs minimal regret, defined as the difference between the total expected rewards accumulated by a model-aware genie, and that obtained by all users applying the policy. We make two contributions in this paper. First, we consider the setting where the users have a prioritized ranking, such that it is desired for the $K$-th-ranked user to learn to access the arm offering the $K$-th highest mean reward. For this problem, we present the first distributed policy that yields regret that is uniformly logarithmic over time without requiring any prior assumption about the mean rewards. Second, we consider the case when a fair access policy is required, i.e., it is desired for all users to experience the same mean reward. For this problem, we present a distributed policy that yields order-optimal regret scaling with respect to the number of users and arms, better than previously proposed policies in the literature. Both of our distributed policies make use of an innovative modification of the well known UCB1 policy for the classic multi-armed bandit problem that allows a single user to learn how to play the arm that yields the $K$-th largest mean reward.

The fundamental problem of multiple secondary users contending for opportunistic spectrum access over multiple channels in cognitive radio networks has been formulated recently as a decentralized multi-armed bandit (D-MAB) problem. In a D-MAB problem there are $M$ users and $N$ arms (channels) that each offer i.i.d. stochastic rewards with unknown means so long as they are accessed without collision. The goal is to design a decentralized online learning policy that incurs minimal regret, defined as the difference between the total expected rewards accumulated by a model-aware genie, and that obtained by all users applying the policy. We make two contributions in this paper. First, we consider the setting where the users have a prioritized ranking, such that it is desired for the $K$-th-ranked user to learn to access the arm offering the $K$-th highest mean reward. For this problem, we present the first distributed policy that yields regret that is uniformly logarithmic over time without requiring any prior assumption about the mean rewards. Second, we consider the case when a fair access policy is required, i.e., it is desired for all users to experience the same mean reward. For this problem, we present a distributed policy that yields order-optimal regret scaling with respect to the number of users and arms, better than previously proposed policies in the literature. Both of our distributed policies make use of an innovative modification of the well known UCB1 policy for the classic multi-armed bandit problem that allows a single user to learn how to play the arm that yields the $K$-th largest mean reward.

Using quantum algorithms to improve deep learning models' capabilities is becoming increasingly popular as quantum computing develops. In this work, we investigate how quantum algorithms using quantum neural networks (QNNs) might enhance the effectiveness and performance of deep learning models. We examine the effects of quantum-inspired methods on tasks, including regression, sorting, and optimization, by thoroughly analyzing quantum algorithms and how they integrate with deep learning systems. We experiment with Estimator QNN and Sampler QNN implementations using Qiskit machine-learning, analyzing their forward and backward pass outcomes to assess the effectiveness of quantum algorithms in improving deep learning models. Our research clarifies the scope, intricacy, and scalability issues surrounding QNNs and offers insights into the possible advantages and difficulties of quantum-enhanced deep learning. This work adds to the continuing investigation of quantum computing's potential to advance machine learning and artificial intelligence paradigms by clarifying the interaction between quantum algorithms and deep learning systems.

Objective: Classifying the risk of policyholders based on observable characteristics can help insurance companies to reduce losses, identify customers more accurately, and prevent adverse selection in the insurance market. The purpose of this article is to examine the financial losses caused by third party insurance and to predict the risk of policyholders in the event of an accident. Methodology: using decision tree algorithms, support vector machine, Naive Bayes and neural network; The hidden data patterns have been discovered in order to classify third party insurance policyholders. Also, the unbalanced distribution of data in two groups of damaged and undamaged causes an important challenge in the application of machine learning and data mining methods, which is considered in this article. Findings: The data set belongs to one of the insurance companies and contains more than four hundred thousand samples registered in five years and includes four independent variables of car type, car group, license plate type and car age and a dependent and two-valued variable of financial damage. According to the obtained results, the best performance and prediction accuracy (with accuracy F1=0.72±0.01) is related to the decision tree model. Conclusion: The impact of variables on the occurrence of damage in order of priority are: car type, license plate type, car age and car group. The evaluation results show that more data related to the driver's characteristics is needed for more accurate prediction of damage and high-risk customers.

Available literature reports several lymphoma cases misdiagnosed as tuberculosis, especially in countries with a heavy TB burden. This frequent misdiagnosis is due to the fact that the two diseases can present with similar symptoms. The present study therefore aims to analyse and explore TB as well as lymphoma case reports using Natural Language Processing tools and evaluate the use of machine learning to differentiate between the two diseases. As a starting point in the study, case reports were collected for each disease using web scraping. Natural language processing tools and text clustering were then used to explore the created dataset. Finally, six machine learning algorithms were trained and tested on the collected data, which contained 765 lymphoma and 546 tuberculosis case reports. Each method was evaluated using various performance metrics. The results indicated that the multi-layer perceptron model achieved the best accuracy (93.1%), recall (91.9%) and precision score (93.7%), thus outperforming other algorithms in terms of correctly classifying the different case reports.

Agriculture and farming are the most important and basic industries that are very important to humanity and generate a considerable portion of any nation's GDP. For good agricultural and farming management, technological advancements and support are requir ed. Smart ag riculture (or) farm ing is a set of approaches that uses a variety of current information and communication technology to improve the production and quality of agricultural products with minimum human involvement and at a lower cost. Smart farmi ng is mostly based on IoT technology, since there is a need to continually monitor numerous aspects in the agricultural field, such as water level, light, soil characteristics, plant development, and so on. Machine learning algorithms are used in smart far ming to increase production and reduce the risk of crop damage. Data analytics has been shown through extensive study to improve the accuracy and predictability of smart agricultural systems. Data analytics is utilised in agricultural fields to make decisi ons and recommend acceptable crops for production. This study provides a comprehensive overview of the different methods and structures utilised in smart farming. It also provides a thorough analysis of different designs and recommends appropriate answers to today's smart farming problems.

Summer monsoon rainfall contributes more than 75% of the annual rainfall in India. For the state of Maharashtra, India, this is more than 80% for almost all regions of the state. The high variability of rainfall during this period necessitates the classification of rainy and non-rainy days. While there are various approaches to rainfall classification, this paper proposes rainfall classification based on weather variables. This paper explores the use of support vector machine (SVM) and artificial neural network (ANN) algorithms for the binary classification of summer monsoon rainfall using common weather variables such as relative humidity, temperature, pressure. The daily data, for the summer monsoon months, for nineteen years, was collected for the Shivajinagar station of Pune in the state of Maharashtra, India. Classification accuracy of 82.1 and 82.8%, respectively, was achieved with SVM and ANN algorithms, for an imbalanced dataset. While performance parameters such as misclassification rate, F1 score indicate that better results were achieved with ANN, model parameter selection for SVM was less involved than ANN. Domain adaptation technique was used for rainfall classification at the other two stations of Maharashtra with the network trained for the Shivajinagar station. Satisfactory results for these two stations were obtained only after changing the training method for SVM and ANN.

A quick and precise crop leaf disease detection is important to increasing agricultural yield in a sustainable manner. We present a comprehensive overview of recent research in the field of crop leaf disease prediction using image processing (IP), machine learning (ML) and deep learning (DL) techniques in this paper. Using these techniques, crop leaf disease prediction made it possible to get notable accuracies. This article presents a survey of research papers that presented the various methodologies, analyzes them in terms of the dataset, number of images, number of classes, algorithms used, convolutional neural networks (CNN) models employed, and overall performance achieved. Then, suggestions are prepared on the most appropriate algorithms to deploy in standard, mobile/embedded systems, drones, robots and unmanned aerial vehicles (UAV). We discussed the performance measures used and listed some of the limitations and future works that requires to be focus on, to extend real time automated crop leaf disease detection system.

A quick and precise crop leaf disease detection is important to increasing agricultural yield in a sustainable manner. We present a comprehensive overview of recent research in the field of crop leaf disease prediction using image processing (IP), machine learning (ML) and deep learning (DL) techniques in this paper. Using these techniques, crop leaf disease prediction made it possible to get notable accuracies. This article presents a survey of research papers that presented the various methodologies, analyzes them in terms of the dataset, number of images, number of classes, algorithms used, convolutional neural networks (CNN) models employed, and overall performance achieved. Then, suggestions are prepared on the most appropriate algorithms to deploy in standard, mobile/embedded systems, drones, robots and unmanned aerial vehicles (UAV). We discussed the performance measures used and listed some of the limitations and future works that requires to be focus on, to extend real time automated crop leaf disease detection system.

In this report we describe a tool for comparing the performance of causal structure learning algorithms implemented in the TETRAD freeware suite of causal analysis methods. Currently the tool is available as a package in the TETRAD source code (written in Java), which can be loaded up in an Integrated Development Environment (IDE) such as IntelliJ IDEA. Simulations can be done varying the number of runs, sample sizes, and data modalities. Performance on this simulated data can then be compared for a number of algorithms, with parameters varied and with performance statistics as selected, producing a publishable report. The order of the algorithms in the output can be adjusted to the user's preference using a utility function over the statistics. Data sets from simulation can be saved along with their graphs to a file and loaded back in for further analysis, or used for analysis by other tools.

A recent goal in the Reinforcement Learning (RL) framework is to choose a sequence of actions or a policy to maximize the reward collected or minimize the regret incurred in a finite time horizon. For several RL problems in operation research and optimal control, the optimal policy of the underlying Markov Decision Process (MDP) is characterized by a known structure. The current state of the art algorithms do not utilize this known structure of the optimal policy while minimizing regret. In this work, we develop new RL algorithms that exploit the structure of the optimal policy to minimize regret. Numerical experiments on MDPs with structured optimal policies show that our algorithms have better performance, are easy to implement, have a smaller run-time and require less number of random number generations.

The purpose of this paper is to introduce a concept of equivalence between machine learning algorithms. We define two notions of algorithmic equivalence, namely, weak and strong equivalence. These notions are of paramount importance for identifying when learning prop erties from one learning algorithm can be transferred to another. Using regularized kernel machines as a case study, we illustrate the importance of the introduced equivalence concept by analyzing the relation between kernel ridge regression (KRR) and m-power regularized least squares regression (M-RLSR) algorithms.

We study the problem of dynamic spectrum sensing and access in cognitive radio systems as a partially observed Markov decision process (POMDP). A group of cognitive users cooperatively tries to exploit vacancies in primary (licensed) channels whose occupancies follow a Markovian evolution. We first consider the scenario where the cognitive users have perfect knowledge of the distribution of the signals they receive from the primary users. For this problem, we obtain a greedy channel selection and access policy that maximizes the instantaneous reward, while satisfying a constraint on the probability of interfering with licensed transmissions. We also derive an analytical universal upper bound on the performance of the optimal policy. Through simulation, we show that our scheme achieves good performance relative to the upper bound and improved performance relative to an existing scheme. We then consider the more practical scenario where the exact distribution of the signal from the primary is unknown. We assume a parametric model for the distribution and develop an algorithm that can learn the true distribution, still guaranteeing the constraint on the interference probability. We show that this algorithm outperforms the naive design that assumes a worst case value for the parameter. We also provide a proof for the convergence of the learning algorithm.

This thesis explores the power of interactivity in unsupervised machine learning problems. Interactive algorithms employ feedback-driven measurements to reduce data acquisition costs and consequently enable statistical analysis in otherwise intractable settings. Unsupervised learning methods are fundamental tools across a variety of domains, and interactive procedures promise to broaden the scope of statistical analysis. We develop interactive learning algorithms for three unsupervised problems: subspace learning, clustering, and tree metric learning. Our theoretical and empirical analysis shows that interactivity can bring both statistical and computational improvements over non-interactive approaches. An over-arching thread of this thesis is that interactive learning is particularly powerful for non-uniform datasets, where non-uniformity is quantified differently in each setting. We first study the subspace learning problem, where the goal is to recover or approximate the principal subspace of a collection of partially observed data points. We propose statistically and computationally appealing interactive algorithms for both the matrix completion problem, where the data points lie on a low dimensional subspace, and the matrix approximation problem, where one must approximate the principal components of a collection of points. We measure uniformity with the notion of incoherence, and we show that our feedback-driven algorithms perform well under much milder incoherence assumptions. We next consider clustering a dataset represented by a partially observed similarity matrix. We propose an interactive procedure for recovering a clustering from a small number of carefully selected similarity measurements. The algorithm exploits non-uniformity of cluster size, using few measurements to recover larger clusters and focusing measurements on the smaller structures.

As the military continues to expand its use of intelligent agents in a variety of operational aspects, event prediction and learning algorithms are becoming more and more important. In this paper, we conduct a detailed analysis of two such algorithms: Variable Order Markov and Look-Up Table models. Each model employs different parameters for prediction and this study attempts to determine which model is more accurate in its prediction and why. We find the models contrast in that the Variable Order Markov Model increases its average prediction probability, our primary performance measure, with increased maximum model order, while the Look-Up Table Model decreases average prediction probability with increased recency time threshold. In addition, statistical tests of results of each model indicate a consistency in each model's prediction capabilities, and most of the variation in the results could be explained by model parameters.

A concise and self-contained introduction to causal inference, increasingly important in data science and machine learning.The mathematization of causality is a relatively recent development, and has become increasingly important in data science and machine learning. This book offers a self-contained and concise introduction to causal models and how to learn them from data. After explaining the need for causal models and discussing some of the principles underlying causal inference, the book teaches readers how to use causal models: how to compute intervention distributions, how to infer causal models from observational and interventional data, and how causal ideas could be exploited for classical machine learning problems. All of these topics are discussed first in terms of two variables and then in the more general multivariate case. The bivariate case turns out to be a particularly hard problem for causal learning because there are no conditional independences as used by classical methods for solving multivariate cases. The authors consider analyzing statistical asymmetries between cause and effect to be highly instructive, and they report on their decade of intensive research into this problem. The book is accessible to readers with a background in machine learning or statistics, and can be used in graduate courses or as a reference for researchers. The text includes code snippets that can be copied and pasted, exercises, and an appendix with a summary of the most important technical concepts.

Preeclampsia and eclampsia are the most common obstetric disorders associated with poor maternal and neonatal outcome. The study’s primary objective is to assess the accuracy of novel high-risk factors core using machine learning algorithms in predicting preeclampsia. The study included 400 pregnant women and used 27 novel high-risk factors to predict preeclampsia. The target variables for predicting preeclampsia are systolic and diastolic blood pressures. Various algorithms, including decision tree (DT), random forest (RF), gradient boosting, support vector machine (SVM), K-neighbors, light gradient boosting machine (LGBM), multi-layer perceptron (MLP), Adaboost classifier, and extra trees classifier are used in the analysis. The accuracy and precision of the LGBM classifier (0.85 and 0.9583 with F1 0.7188), support vector classifier (0.8417 and 0.92 with F1 0.7077), DT (0.825 and 0.913 with F1 0.6667), and extra trees (0.8167 and 0.9091 with F1 0.6452) are found to be better algorithms for prediction of preeclampsia. According to the novel high-risk factors score, 17.5% of pregnant women were identified as being at high risk for preeclampsia during the first trimester, which increased to 18.7% in 3rd trimester; in addition, 16% of pregnant women had a blood pressure of 140/90 mmHg and the above. Novel, high-risk scores and machine learning algorithms can effectively predict preeclampsia at an early period.

In this paper we explore the use of an adaptive search technique (genetic algorithms) to construct a system GABIL which continually learns and refines concept classification rules from its interaction with the environment. The performance of the system is measured on a set of concept learning problems and compared with the performance of two existing systems: ID5R and C4.5. Preliminary results support that, despite minimal system bias, GABIL is an effective concept learner and is quite competitive with ID5R and C4.5 as the target concept increases in complexity. (AN)

We consider the problem of finding stationary Nash equilibria (NE) in a finite discounted general-sum stochastic game. We first generalize a non-linear optimization problem from Filar and Vrieze [2004] to a $N$-player setting and break down this problem into simpler sub-problems that ensure there is no Bellman error for a given state and an agent. We then provide a characterization of solution points of these sub-problems that correspond to Nash equilibria of the underlying game and for this purpose, we derive a set of necessary and sufficient SG-SP (Stochastic Game - Sub-Problem) conditions. Using these conditions, we develop two actor-critic algorithms: OFF-SGSP (model-based) and ON-SGSP (model-free). Both algorithms use a critic that estimates the value function for a fixed policy and an actor that performs descent in the policy space using a descent direction that avoids local minima. We establish that both algorithms converge, in self-play, to the equilibria of a certain ordinary differential equation (ODE), whose stable limit points coincide with stationary NE of the underlying general-sum stochastic game. On a single state non-generic game (see Hart and Mas-Colell [2005]) as well as on a synthetic two-player game setup with $810,000$ states, we establish that ON-SGSP consistently outperforms NashQ ([Hu and Wellman, 2003] and FFQ [Littman, 2001] algorithms.

We study quantum learning algorithms for quantum measurements. The optimal learning algorithm is derived for arbitrary von Neumann measurements in the case of training with one or two examples. The analysis of the case of three examples reveals that, differently from the learning of unitary gates, the optimal algorithm for learning of quantum measurements cannot be parallelized, and requires quantum memories for the storage of information.

This 60th episode of Learning Machines 101 discusses how one can use novelty detection or anomaly detection machine learning algorithms to monitor the performance of other machine learning algorithms deployed in real world environments. The episode is based upon a review of a talk by Chief Data Scientist Ira Cohen of Anodot presented at the 2016 Berlin Buzzwords Data Science Conference. Check out: www.learningmachines101.com to hear the podcast or read a transcription of the podcast!

Learning an algorithm--a systematic sequence of operations for solving a problem with given input--is often difficult for students due to the abstract nature of the algorithms and the data they process. To help students understand the behavior of algorithms, a subfield in computing education research has focused on algorithm visualization--learning material showing the steps and data used by an algorithm. As the use of mobile devices has risen together with the capabilities of the devices, mobile learning is more important than ever. It also opens possibilities to contextualize the learning experience. In this paper, we present our work towards location-aware mobile learning of spatial algorithms that adapts the learning material to the location of the student. [For the full proceedings see ED562140.]

Although the advantage of computational pathology is vast it does have challenges. Thereby, aim to evaluate the deep learning methods employed for digitized WSI histopathology image classification in cancers. Also, to elicit the challenges while employing DL and provide solutions specific to WSI digitalized pathology image analysis in cancers.