Algorithms, Volume 15, Issue 5 (May 2022) – 41 articles

The human immunodeficiency virus (HIV) mainly attacks CD4⁺ T cells in the host. Chronic HIV infection gradually depletes the CD4⁺ T cell pool, compromising the host’s immunological reaction to invasive infections and ultimately leading to acquired immunodeficiency syndrome (AIDS). The goal of this study is not to provide a qualitative description of the rich dynamic characteristics of the HIV infection model of CD4⁺ T cells, but to produce accurate analytical solutions to the model using the modified iterative approach. In this research, a new efficient method using the new iterative method (NIM), the coupling of the standard NIM and Laplace transform, called the modified new iterative method (MNIM), has been introduced to resolve the HIV infection model as a class of system of ordinary differential equations (ODEs). A nonlinear HIV infection dynamics model is adopted as an instance to elucidate the identification process and the solution process of MNIM, only two iterations lead to ideal results. In addition, the model has also been solved using NIM and the fourth order Runge–Kutta (RK4) method. The results indicate that the solutions by MNIM match with those of RK4 method to a minimum of eight decimal places, whereas NIM solutions are not accurate enough. Numerical comparisons between the MNIM, NIM, the classical RK4 and other methods reveal that the modified technique has potential as a tool for the nonlinear systems of ODEs. Full article

Concept drift, which refers to changes in the underlying process structure or customer behaviour over time, is inevitable in business processes, causing challenges in ensuring that the learned model is a proper representation of the new data. Due to factors such as seasonal effects and policy updates, concept drifts can occur in customer transitions and time spent throughout the process, either suddenly or gradually. In a concept drift context, we can discard the old data and retrain the model using new observations (sudden drift) or combine the old data with the new data to update the model (gradual drift) or maintain the model as unchanged (no drift). In this paper, we model a response to concept drift as a sequential decision making problem by combing a hierarchical Markov model and a Markov decision process (MDP). The approach can detect concept drift, retrain the model and update customer profiles automatically. We validate the proposed approach on 68 artificial datasets and a real-world hospital billing dataset, with experimental results showing promising performance. Full article

Construction companies are increasingly utilizing sensing technologies to automatically record different steps of the construction process in detail for effective monitoring and control. This generates a significant amount of event data that can be used to learn the underlying behavior of agents in a construction site using process mining. While process mining can be used to discover the real process and identify and analyze deviations and bottlenecks in operations, it is a backward-looking approach. On the other hand, discrete event simulation (DES) provides a means to forecast future performance from historical data to enable proactive decision-making by operation managers relating to their projects. However, this method is largely unused by the industry due to the specialized knowledge required to create the DES models. This paper thus proposes a framework that extends the utility of collecting event data and their process models, by transforming them into DES models for forecasting future performance. This framework also addresses another challenge of using DES relating to its inability to update itself as the project progresses. This challenge is addressed by using the Bayesian updating technique to continuously update the input parameters of the simulation model for the most up-to-date estimation based on data collected from the field. The proposed framework was validated on a real-world case study of an earthmoving operation. The results show that the process mining techniques could accurately discover the process model from the event data collected from the field. Furthermore, it was noted that continuous updating of DES model input parameters can provide accurate and reliable productivity estimates based on the actual data generated from the field. The proposed framework can help stakeholders to discover the underlying sequence of their operations, and enable timely, data-driven decisions regarding operations control. Full article

Considering the characteristics of different types of users in hybrid carsharing systems, in which sharing autonomous vehicles (SAVs) and conventional sharing cars (CSCs) coexist, a tailored pricing strategy (TPS) is proposed to maximize the operator’s profit and minimize all users’ costs. The fleet sizes and sizes of SAVs’ stations are also determined simultaneously. A bi-objective nonlinear programming model is established, and a genetic algorithm is applied to solve it. Based on the operational data in Lanzhou, China, carsharing users are clustered into three types. They are loyal users, losing users, and potential users, respectively. Results show the application of the TPS can help the operator increase profit and attract more users. The loyal users are assigned the highest price, while they still contribute the most to the operator’s profit with the highest number of carsharing trips. The losing users and potential users are comparable in terms of the number of trips, while the latter still makes more profit. Full article

In order to obtain the optimal perspectives of the recognition target, this paper combines the motion path of the manipulator arm and camera. A path planning method to find the optimal perspectives based on an A* algorithm is proposed. The quality of perspectives is represented by means of multi-view recognition. A binary multi-view 2D kernel principal component analysis network (BM2DKPCANet) is built to extract features. The multi-view angles classifier based on BM2DKPCANet + Softmax is established, which outputs category posterior probability to represent the perspective recognition performance function. The path planning problem is transformed into a multi-objective optimization problem by taking the optimal view recognition and the shortest path distance as the objective functions. In order to reduce the calculation, the multi-objective optimization problem is transformed into a single optimization problem by fusing the objective functions based on the established perspective observation directed graph model. An A* algorithm is used to solve the single source shortest path problem of the fused directed graph. The path planning experiments with different numbers of view angles and different starting points demonstrate that the method can guide the camera to reach the viewpoint with higher recognition accuracy and complete the optimal observation path planning. Full article

Clustering is a critical part of many tasks and, in most applications, the number of clusters in the data are unknown and must be estimated. This paper presents an Extreme Value Theory-based approach to threshold selection for clustering, proving that the “correct” linkage distances must follow a Weibull distribution for smooth feature spaces. Deep networks and their associated deep features have transformed many aspects of learning, and this paper shows they are consistent with our extreme-linkage theory and provide Unreasonable Clusterability. We show how our novel threshold selection can be applied to both classic agglomerative clustering and the more recent FINCH (First Integer Neighbor Clustering Hierarchy) algorithm. Our evaluation utilizes over a dozen different large-scale vision datasets/subsets, including multiple face-clustering datasets and ImageNet for both in-domain and, more importantly, out-of-domain object clustering. Across multiple deep features clustering tasks with very different characteristics, our novel automated threshold selection performs well, often outperforming state-of-the-art clustering techniques even when they select parameters on the test set. Full article

Transcranial direct current stimulation (tDCS) is a non-invasive neuromodulation technique to treat brain disorders by using a constant, low current to stimulate targeted cortex regions. Compared to the conventional tDCS that uses two large pad electrodes, multiple electrode tDCS has recently received more attention. It is able to achieve better stimulation performance in terms of stimulation intensity and focality. In this paper, we first establish a computational model of tDCS, and then propose a novel optimization algorithm using a regularization matrix $\lambda$ to explore the balance between stimulation intensity and focality. The simulation study is designed such that the performance of state-of-the-art algorithms and the proposed algorithm can be compared via quantitative evaluation. The results show that the proposed algorithm not only achieves desired intensity, but also smaller target error and better focality. Robustness analysis indicates that the results are stable within the ranges of scalp and cerebrospinal fluid (CSF) conductivities, while the skull conductivity is most sensitive and should be carefully considered in real clinical applications. Full article

Model optimization in neuroscience has focused on inferring intracellular parameters from time series observations of the membrane voltage and calcium concentrations. These parameters constitute the fingerprints of ion channel subtypes and may identify ion channel mutations from observed changes in electrical activity. A central question in neuroscience is whether computational methods may obtain ion channel parameters with sufficient consistency and accuracy to provide new information on the underlying biology. Finding single-valued solutions in particular, remains an outstanding theoretical challenge. This note reviews recent progress in the field. It first covers well-posed problems and describes the conditions that the model and data need to meet to warrant the recovery of all the original parameters—even in the presence of noise. The main challenge is model error, which reflects our lack of knowledge of exact equations. We report on strategies that have been partially successful at inferring the parameters of rodent and songbird neurons, when model error is sufficiently small for accurate predictions to be made irrespective of stimulation. Full article

Life-cycle population follow-up data collection is time-consuming and often takes decades. General cohort data studies collect short-to-medium-term data from populations of different age groups. The purpose of constructing a life-cycle simulation method is to find an efficient and reliable way to achieve the way to characterize life-cycle disease metastasis from these short-to-medium-term data. In this paper, we have presented our effort at construction of a full lifetime population cohort simulation framework. The design aim is to generate a comprehensive understanding of the disease transition for full lifetime when we only have short-or-medium term population cohort data. We have conducted several groups of experiments to show the effectiveness of our method. Full article

Autism spectrum disorder (ASD) is a neurodevelopmental disorder that severely impairs an individual’s cognitive, linguistic, object recognition, communication, and social abilities. This situation is not treatable, although early detection of ASD can assist to diagnose and take proper steps for mitigating its effect. Using various artificial intelligence (AI) techniques, ASD can be detected an at earlier stage than with traditional methods. The aim of this study was to propose a machine learning model that investigates ASD data of different age levels and to identify ASD more accurately. In this work, we gathered ASD datasets of toddlers, children, adolescents, and adults and used several feature selection techniques. Then, different classifiers were applied into these datasets, and we assessed their performance with evaluation metrics including predictive accuracy, kappa statistics, the f1-measure, and AUROC. In addition, we analyzed the performance of individual classifiers using a non-parametric statistical significant test. For the toddler, child, adolescent, and adult datasets, we found that Support Vector Machine (SVM) performed better than other classifiers where we gained 97.82% accuracy for the RIPPER-based toddler subset; 99.61% accuracy for the Correlation-based feature selection (CFS) and Boruta CFS intersect (BIC) method-based child subset; 95.87% accuracy for the Boruta-based adolescent subset; and 96.82% accuracy for the CFS-based adult subset. Then, we applied the Shapley Additive Explanations (SHAP) method into different feature subsets, which gained the highest accuracy and ranked their features based on the analysis. Full article

Data augmentation methods enrich datasets with augmented data to improve the performance of neural networks. Recently, automated data augmentation methods have emerged, which automatically design augmentation strategies. The existing work focuses on image classification and object detection, whereas we provide the first study on semantic image segmentation and introduce two new approaches: SmartAugment and SmartSamplingAugment. SmartAugment uses Bayesian Optimization to search a rich space of augmentation strategies and achieves new state-of-the-art performance in all semantic segmentation tasks we consider. SmartSamplingAugment, a simple parameter-free approach with a fixed augmentation strategy, competes in performance with the existing resource-intensive approaches and outperforms cheap state-of-the-art data augmentation methods. Furthermore, we analyze the impact, interaction, and importance of data augmentation hyperparameters and perform ablation studies, which confirm our design choices behind SmartAugment and SmartSamplingAugment. Lastly, we will provide our source code for reproducibility and to facilitate further research. Full article

The general problem of tiling finite regions of the plane with polyominoes is $\mathcal{NP}$-complete, and so the associated computational geometry problem rapidly becomes intractable for large instances. Thus, the need to reduce algorithm complexity for tiling is important and continues as a fruitful area of research. Traditional approaches to tiling with polyominoes use backtracking, which is a refinement of the ‘brute-force’ solution procedure for exhaustively finding all solutions to a combinatorial search problem. In this work, we combine checkerboard colouring techniques with a recently introduced integer linear programming (ILP) technique for tiling with polyominoes. The colouring arguments often split large tiling problems into smaller subproblems, each represented as a separate ILP problem. Problems that are amenable to this approach are embarrassingly parallel, and our work provides proof of concept of a parallelizable algorithm. The main goal is to analyze when this approach yields a potential parallel speedup. The novel colouring technique shows excellent promise in yielding a parallel speedup for finding large tiling solutions with ILP, particularly when we seek a single (optimal) solution. We also classify the tiling problems that result from applying our colouring technique according to different criteria and compute representative examples using a combination of MATLAB and CPLEX, a commercial optimization package that can solve ILP problems. The collections of MATLAB programs PARIOMINOES (v3.0.0) and POLYOMINOES (v2.1.4) used to construct the ILP problems are freely available for download. Full article

Although graph theory has already been introduced in spatial reasoning, current spatial database systems do not provide out-of-the-box routing on geometric points that are not matched on the graph. Methods that connect new reference locations to the graph render different routing results. Moreover, current solutions break reasoning down to local analysis. We bridge the gap between routing networks and spatial geometry by a global matching of geometric points to routing networks. Full article

A reliable quantification of the worst-case influence of model uncertainty and external disturbances is crucial for the localization of vessels in marine applications. This is especially true if uncertain GPS-based position measurements are used to update predicted vessel locations that are obtained from the evaluation of a ship’s state equation. To reflect real-life working conditions, these state equations need to account for uncertainty in the system model, such as imperfect actuation and external disturbances due to effects such as wind and currents. As an application scenario, the GPS-based localization of autonomous DDboat robots is considered in this paper. Using experimental data, the efficiency of an ellipsoidal approach, which exploits a bounded-error representation of disturbances and uncertainties, is demonstrated. Full article

The finger vein recognition system uses blood vessels inside the finger of an individual for identity verification. The public is in favor of a finger vein recognition system over conventional passwords or ID cards as the biometric technology is harder to forge, misplace, and share. In this study, the histogram of oriented gradients (HOG) features, which are robust against changes in illumination and position, are extracted from the finger vein for personal recognition. To further increase the amount of information that can be used for recognition, different instances of the finger vein, ranging from the index, middle, and ring finger are combined to form a multi-instance finger vein representation. This fusion approach is preferred since it can be performed without requiring additional sensors or feature extractors. To combine different instances of finger vein effectively, score level fusion is adopted to allow greater compatibility among the wide range of matches. Towards this end, two methods are proposed: Bayesian optimized support vector machine (SVM) score fusion (BSSF) and Bayesian optimized SVM based fusion (BSBF). The fusion results are incrementally improved by optimizing the hyperparameters of the HOG feature, SVM matcher, and the weighted sum of score level fusion using the Bayesian optimization approach. This is considered a kind of knowledge-based approach that takes into account the previous optimization attempts or trials to determine the next optimization trial, making it an efficient optimizer. By using stratified cross-validation in the training process, the proposed method is able to achieve the lowest EER of 0.48% and 0.22% for the SDUMLA-HMT dataset and UTFVP dataset, respectively. Full article

Large-scale automatic speech recognition model has achieved impressive performance. However, huge computational resources and massive amount of data are required to train an ASR model. Knowledge distillation is a prevalent model compression method which transfers the knowledge from large model to small model. To improve the efficiency of knowledge distillation for end-to-end speech recognition especially in the low-resource setting, a Mixup-based Knowledge Distillation (MKD) method is proposed which combines Mixup, a data-agnostic data augmentation method, with softmax-level knowledge distillation. A loss-level mixture is presented to address the problem caused by the non-linearity of label in the KL-divergence when adopting Mixup to the teacher–student framework. It is mathematically shown that optimizing the mixture of loss function is equivalent to optimize an upper bound of the original knowledge distillation loss. The proposed MKD takes the advantage of Mixup and brings robustness to the model even with a small amount of training data. The experiments on Aishell-1 show that MKD obtains a 15.6% and 3.3% relative improvement on two student models with different parameter scales compared with the existing methods. Experiments on data efficiency demonstrate MKD achieves similar results with only half of the original dataset. Full article

Scientific papers published in journals or conferences, also considered academic publications, are the manifestation of scientific research achievements. Lots of scientific papers published in digital form bring new challenges for academic evaluation and information retrieval. Therefore, research on the ranking method of scientific papers is significant for the management and evaluation of academic resources. In this paper, we first identify internal and external factors for evaluating scientific papers and propose a publication ranking method based on an analysis of a heterogeneous academic network. We use four types of metadata (i.e., author, venue (journal or conference), topic, and title) as vertexes for creating the network; in there, the topics are trained by the SentenceLDA algorithm with the metadata of the abstract. We then use the Gibbs sampling method to create a heterogeneous academic network and apply the ConNetClus algorithm to calculate the probability value of publication ranking. To evaluate the significance of the method proposed in this paper, we compare the ranking results with BM25, PageRank, etc., and homogeneous networks in MAP and NDCG. As shown in our evaluation results, the performance of the method we propose in this paper is better than other baselines for ranking publications. Full article

A high-quality and secure touchdown run for an aircraft is essential for economic, operational, and strategic reasons. The shortest viable touchdown run without any skidding requires variable braking pressure to manage the friction between the road surface and braking tire at all times. Therefore, the manipulation and regulation of the anti-skid braking system (ABS) should be able to handle steady nonlinearity and undetectable disturbances and to regulate the wheel slip ratio to make sure that the braking system operates securely. This work proposes an active disturbance rejection control technique for the anti-skid braking system. The control law ensures action that is bounded and manageable, and the manipulating algorithm can ensure that the closed-loop machine works around the height factor of the secure area of the friction curve, thereby improving overall braking performance and safety. The stability of the proposed algorithm is proven primarily by means of Lyapunov-based strategies, and its effectiveness is assessed by means of simulations on a semi-physical aircraft brake simulation platform. Full article

A new hybrid metaheuristic method for optimizing the objective function on a parallelepiped set of admissible solutions is proposed. It mimics the behavior of a school of river perch when looking for food. The algorithm uses the ideas of several methods: a frog-leaping method, migration algorithms, a cuckoo algorithm and a path-relinking procedure. As an application, a wide class of problems of finding the optimal control of deterministic discrete dynamical systems with a nonseparable performance criterion is chosen. For this class of optimization problems, it is difficult to apply the discrete maximum principle and its generalizations as a necessary optimality condition and the Bellman equation as a sufficient optimality condition. The desire to extend the class of problems to be solved to control problems of trajectory bundles and stochastic problems leads to the need to use not only classical adaptive random search procedures, but also new approaches combining the ideas of migration algorithms and swarm intelligence methods. The efficiency of this method is demonstrated and an analysis is performed by solving several optimal deterministic discrete control problems: two nonseparable problems (Luus–Tassone and LiHaimes) and five classic linear systems control problems with known exact solutions. Full article

The bio-inspired research field has evolved greatly in the last few years due to the large number of novel proposed algorithms and their applications. The sources of inspiration for these novel bio-inspired algorithms are various, ranging from the behavior of groups of animals to the properties of various plants. One problem is the lack of one bio-inspired algorithm which can produce the best global solution for all types of optimization problems. The presented solution considers the proposal of a novel approach for feature selection in classification problems, which is based on a binary version of a novel bio-inspired algorithm. The principal contributions of this article are: (1) the presentation of the main steps of the original Horse Optimization Algorithm (HOA), (2) the adaptation of the HOA to a binary version called the Binary Horse Optimization Algorithm (BHOA), (3) the application of the BHOA in feature selection using nine state-of-the-art datasets from the UCI machine learning repository and the classifiers Random Forest (RF), Support Vector Machines (SVM), Gradient Boosted Trees (GBT), Logistic Regression (LR), K-Nearest Neighbors (K-NN), and Naïve Bayes (NB), and (4) the comparison of the results with the ones obtained using the Binary Grey Wolf Optimizer (BGWO), Binary Particle Swarm Optimization (BPSO), and Binary Crow Search Algorithm (BCSA). The experiments show that the BHOA is effective and robust, as it returned the best mean accuracy value and the best accuracy value for four and seven datasets, respectively, compared to BGWO, BPSO, and BCSA, which returned the best mean accuracy value for four, two, and two datasets, respectively, and the best accuracy value for eight, seven, and five datasets, respectively. Full article

A number of AI-generated tools are used today to clone human voices, leading to a new technology known as Audio Deepfakes (ADs). Despite being introduced to enhance human lives as audiobooks, ADs have been used to disrupt public safety. ADs have thus recently come to the attention of researchers, with Machine Learning (ML) and Deep Learning (DL) methods being developed to detect them. In this article, a review of existing AD detection methods was conducted, along with a comparative description of the available faked audio datasets. The article introduces types of AD attacks and then outlines and analyzes the detection methods and datasets for imitation- and synthetic-based Deepfakes. To the best of the authors’ knowledge, this is the first review targeting imitated and synthetically generated audio detection methods. The similarities and differences of AD detection methods are summarized by providing a quantitative comparison that finds that the method type affects the performance more than the audio features themselves, in which a substantial tradeoff between the accuracy and scalability exists. Moreover, at the end of this article, the potential research directions and challenges of Deepfake detection methods are discussed to discover that, even though AD detection is an active area of research, further research is still needed to address the existing gaps. This article can be a starting point for researchers to understand the current state of the AD literature and investigate more robust detection models that can detect fakeness even if the target audio contains accented voices or real-world noises. Full article

In this study, the modification of the quantum multi-swarm optimization algorithm is proposed for dynamic optimization problems. The modification implies using the search operators from differential evolution algorithm with a certain probability within particle swarm optimization to improve the algorithm’s search capabilities in dynamically changing environments. For algorithm testing, the Generalized Moving Peaks Benchmark was used. The experiments were performed for four benchmark settings, and the sensitivity analysis to the main parameters of algorithms is performed. It is shown that applying the mutation operator from differential evolution to the personal best positions of the particles allows for improving the algorithm performance. Full article

An acknowledgment of feedback is extremely helpful in medical training, as it may improve student skill development and provide accurate, unbiased feedback. Data are generated by hundreds of complicated and variable processes within healthcare including treatments, lab results, and internal logistics. Additionally, it is crucial to analyze medical training data to improve operational processes and eliminate bottlenecks. Therefore, the use of process mining (PM) along with conformance checking allows healthcare trainees to gain knowledge about instructor training. Researchers find it challenging to analyze the conformance between observations from event logs and predictions from models with artifacts from the training process. To address this conformance check, we modeled student activities and performance patterns in the training of Central Venous Catheter (CVC) installation. This work aims to provide medical trainees with activities with easy and interpretable outcomes. The two independent techniques for mining process models were fuzzy (i.e., for visualizing major activities) and inductive (i.e., for conformance checking at low threshold noise levels). A set of 20 discrete activity traces was used to validate conformance checks. Results show that 97.8% of the fitness of the model and the movement of the model occurred among the nine activities. Full article

In this paper, we quantify the non-transitivity in chess using human game data. Specifically, we perform non-transitivity quantification in two ways—Nash clustering and counting the number of rock–paper–scissor cycles—on over one billion matches from the Lichess and FICS databases. Our findings indicate that the strategy space of real-world chess strategies has a spinning top geometry and that there exists a strong connection between the degree of non-transitivity and the progression of a chess player’s rating. Particularly, high degrees of non-transitivity tend to prevent human players from making progress in their Elo ratings. We also investigate the implications of non-transitivity for population-based training methods. By considering fixed-memory fictitious play as a proxy, we conclude that maintaining large and diverse populations of strategies is imperative to training effective AI agents for solving chess. Full article

Recent developments have shown that the widely used simplified differential model of Eringen’s nonlocal elasticity in nanobeam analysis is not equivalent to the corresponding and initially proposed integral models, the pure integral model and the two-phase integral model, in all cases of loading and boundary conditions. This has resolved a paradox with solutions that are not in line with the expected softening effect of the nonlocal theory that appears in all other cases. In addition, it revived interest in the integral model and the two-phase integral model, which were not used due to their complexity in solving the relevant integral and integro-differential equations, respectively. In this article, we use a direct operator method for solving boundary value problems for nth order linear Volterra–Fredholm integro-differential equations of convolution type to construct closed-form solutions to the two-phase integral model of Euler–Bernoulli nanobeams in bending under transverse distributed load and various types of boundary conditions. Full article

It is not an easy task for organizers to observe the engagement level of a video meeting audience. This research was conducted to build an intelligent system to enhance the experience of video conversations such as virtual meetings and online classrooms using convolutional neural network (CNN)- and support vector machine (SVM)-based machine learning models to classify the emotional states and the attention level of the participants to a video conversation. This application visualizes their attention and emotion analytics in a meaningful manner. This proposed system provides an artificial intelligence (AI)-powered analytics system with optimized machine learning models to monitor the audience and prepare insightful reports on the basis of participants’ facial features throughout the video conversation. One of the main objectives of this research is to utilize the neural accelerator chip to enhance emotion and attention detection tasks. A custom CNN developed by Gyrfalcon Technology Inc (GTI) named GnetDet was used in this system to run the trained model on their GTI Lightspeeur 2803 neural accelerator chip. Full article

Credit scoring is an effective tool for banks and lending companies to manage the potential credit risk of borrowers. Machine learning algorithms have made grand progress in automatic and accurate discrimination of good and bad borrowers. Notably, ensemble approaches are a group of powerful tools to enhance the performance of credit scoring. Random forest (RF) and Gradient Boosting Decision Tree (GBDT) have become the mainstream ensemble methods for precise credit scoring. RF is a Bagging-based ensemble that realizes accurate credit scoring enriches the diversity base learners by modifying the training object. However, the optimization pattern that works on invariant training targets may increase the statistical independence of base learners. GBDT is a boosting-based ensemble approach that reduces the credit scoring error by iteratively changing the training target while keeping the training features unchanged. This may harm the diversity of base learners. In this study, we incorporate the advantages of the Bagging ensemble training strategy and boosting ensemble optimization pattern to enhance the diversity of base learners. An extreme learning machine-based supervised augmented GBDT is proposed to enhance the discriminative ability for credit scoring. Experimental results on 4 public credit datasets show a significant improvement in credit scoring and suggest that the proposed method is a good solution to realize accurate credit scoring. Full article

Granular Computing is a powerful information processing paradigm, particularly useful for the synthesis of pattern recognition systems in structured domains (e.g., graphs or sequences). According to this paradigm, granules of information play the pivotal role of describing the underlying (possibly complex) process, starting from the available data. Under a pattern recognition viewpoint, granules of information can be exploited for the synthesis of semantically sound embedding spaces, where common supervised or unsupervised problems can be solved via standard machine learning algorithms. In this work, we show a comparison between different strategies for the automatic synthesis of information granules in the context of graph classification. These strategies mainly differ on the specific topology adopted for subgraphs considered as candidate information granules and the possibility of using or neglecting the ground-truth class labels in the granulation process. Computational results on 10 different open-access datasets show that by using a class-aware granulation, performances tend to improve (regardless of the information granules topology), counterbalanced by a possibly higher number of information granules. Full article

In many real-life problems, it is difficult to acquire or label large amounts of data, resulting in so-called few-shot learning problems. However, few-shot classification is a challenging problem due to the uncertainty caused by using few labeled samples. In the past few years, many methods have been proposed with the common aim of transferring knowledge acquired on a previously solved task, which is often achieved by using a pretrained feature extractor. As such, if the initial task contains many labeled samples, it is possible to circumvent the limitations of few-shot learning. A shortcoming of existing methods is that they often require priors about the data distribution, such as the balance between considered classes. In this paper, we propose a novel transfer-based method with a double aim: providing state-of-the-art performance, as reported on standardized datasets in the field of few-shot learning, while not requiring such restrictive priors. Our methodology is able to cope with both inductive cases, where prediction is performed on test samples independently from each other, and transductive cases, where a joint (batch) prediction is performed. Full article

This paper studies the problem of identifying piano music in various modalities using a single, unified approach called marketplace fingerprinting. The key defining characteristic of marketplace fingerprinting is choice: we consider a broad range of fingerprint designs based on a generalization of standard n-grams, and then select the fingerprint designs at runtime that are best for a specific query. We show that the large-scale retrieval problem can be framed as an economics problem in which a consumer and a store interact. In our analogy, the runtime search is like a consumer shopping in the store, the items for sale correspond to fingerprints, and purchasing an item corresponds to doing a fingerprint lookup in the database. Using basic principles of economics, we design an efficient marketplace in which the consumer has many options and adopts a rational buying strategy that explicitly considers the cost and expected utility of each item. We evaluate our marketplace fingerprinting approach on four different sheet music retrieval tasks involving sheet music images, MIDI files, and audio recordings. Using a database containing approximately 375,000 pages of sheet music, our method is able to achieve $0.91$ mean reciprocal rank with sub-second average runtime on cell phone image queries. On all four retrieval tasks, the marketplace method substantially outperforms previous methods while simultaneously reducing average runtime. We present comprehensive experimental results, as well as detailed analyses to provide deeper intuition into system behavior. Full article