Genetic CFL: Hyperparameter Optimization in Clustered Federated Learning

Federated learning (FL) is a distributed model for deep learning that integrates client-server architecture, edge computing, and real-time intelligence. FL has the capability of revolutionizing machine learning (ML) but lacks in the practicality of implementation due to technological limitations, communication overhead, non-IID (independent and identically distributed) data, and privacy concerns. Training a ML model over heterogeneous non-IID data highly degrades the convergence rate and performance. The existing traditional and clustered FL algorithms exhibit two main limitations, including inefficient client training and static hyperparameter utilization. To overcome these limitations, we propose a novel hybrid algorithm, namely, genetic clustered FL (Genetic CFL), that clusters edge devices based on the training hyperparameters and genetically modifies the parameters clusterwise. Then, we introduce an algorithm that drastically increases the individual cluster accuracy by integrating the density-based clustering and genetic hyperparameter optimization. The results are bench-marked using MNIST handwritten digit dataset and the CIFAR-10 dataset. The proposed genetic CFL shows significant improvements and works well with realistic cases of non-IID and ambiguous data. An accuracy of 99.79% is observed in the MNIST dataset and 76.88% in CIFAR-10 dataset with only 10 training rounds.

indelPost: harmonizing ambiguities in simple and complex indel alignments

Summary Small insertions and deletions (indels) in nucleotide sequence may be represented differently between mapping algorithms and variant callers, or in the flanking sequence context. Representational ambiguity is especially profound for complex indels, complicating comparisons between multiple mappings and call sets. Complex indels may additionally suffer from incomplete allele representation, potentially leading to critical misannotation of variant effect. We present indelPost, a Python library that harmonizes these ambiguities for simple and complex indels via realignment and read-based phasing. We demonstrate that indelPost enables accurate analysis of ambiguous data and can derive the correct complex indel alleles from the simple indel predictions provided by standard small variant detectors, with improved performance over a specialized tool for complex indel analysis. Availability indelPost is freely available at: https://github.com/stjude/indelPost. Supplementary information Supplementary data are available at Bioinformatics online.

Two Solitudes: Wilder Penfield, Ewen Cameron, and the Search for a Better Lobotomy

In the 1940s, Wilder Penfield carried out a series of experimental psychosurgeries with the psychiatrist D. Ewen Cameron. This article explores Penfield’s brief foray into psychosurgery and uses this episode to re-examine the emergence of his surgical enterprise. Penfield’s greatest achievement – the surgical treatment of epilepsy – grew from the same roots as psychosurgery, and the histories of these treatments overlap in surprising ways. Within the contexts of Rockefeller-funded neuropsychiatry and Adolf Meyer’s psychobiology, Penfield’s frontal lobe operations (including a key operation on his sister) played a crucial role in the development of lobotomy in the 1930s. The combination of ambiguous data and the desire to collaborate with a psychiatrist encouraged Penfield to try to develop a superior operation. However, unlike his collaboration with psychiatrists, Penfield’s productive working relationship with psychologists encouraged him to abandon the experimental “gyrectomy” procedure. The story of Penfield’s attempt to find a better lobotomy can help us to examine different forms of interdisciplinarity within biomedicine.

Domination in Join of Fuzzy Incidence Graphs Using Strong Pairs with Application in Trading System of Different Countries

Fuzzy graphs (FGs), broadly known as fuzzy incidence graphs (FIGs), are an applicable and well-organized tool to epitomize and resolve multiple real-world problems in which ambiguous data and information are essential. In this article, we extend the idea of domination of FGs to the FIG using strong pairs. An idea of strong pair dominating set and a strong pair domination number (SPDN) is explained with various examples. A theorem to compute SPDN for a complete fuzzy incidence graph (CFIG) is also provided. It is also proved that in any fuzzy incidence cycle (FIC) with l vertices the minimum number of elements in a strong pair dominating set are M[γs(Cl(σ,ϕ,η))]=⌈l3⌉. We define the joining of two FIGs and present a way to compute SPDN in the join of FIGs. A theorem to calculate SPDN in the joining of two strong fuzzy incidence graphs is also provided. An innovative idea of accurate domination of FIGs is also proposed. Some instrumental and useful results of accurate domination for FIC are also obtained. In the end, a real-life application of SPDN to find which country/countries has/have the best trade policies among different countries is examined. Our proposed method is symmetrical to the optimization.

Diagnosis and Management of Testosterone Deficiency in men: A review of the European and American Urology Associations

Ambiguous data on both terminology, diagnostics, and treatment of testosterone deficiency in men prompted us to attempt a critical analysis of existing knowledge on this subject. Current guidelines of both American and European Association of Urology (AUA, EUA) define testosterone therapy as effective and safe. However, media reports and some scientific reports indicating negative effects of the abovementioned therapy arouse aversion to its use by doctors and potential patients for fear of developing prostate cancer or cardiovascular incidents. The peak of scepticism about testosterone therapy was observed after the publication in 2013 and 2014, respectively, two retrospective data analysis on this topic, which resulted in the discontinuation of therapy in many patients with long histories of benefits from testosterone therapy. In addition, in many men with indications for testosterone therapy, this treatment was not used for fear of patient safety. However, the latest data on these concerns do not confirm any negative effects. More than 100 recently published scientific studies have shown the beneficial effects of testosterone therapy on many aspects of health. The American Society of Clinical Endocrinologists (AACE) and the American College of Endocrinology (ACE) have jointly developed their own literature assessment, stating that there is no convincing evidence that testosterone therapy increases the risk of cardiovascular incidents. The same conclusions can be drawn from the current EAU and AUA guidelines.

A Critical Review of Morphologic Findings and Data From 14 Toxicological Studies Involving Fish Exposures to Diclofenac

A number of studies have investigated the potential toxicity of the analgesic agent diclofenac (DCF) in various fish species under a diverse array of experimental conditions. Reported evidence of toxicity in these investigations is often strongly reliant on morphologic end points such as histopathology, immunohistochemistry, and transmission electron microscopy. However, it may be challenging for scientists who perform environmental hazard or risk determination to fully appreciate the intricacies of these specialized endpoints. Therefore, the purpose of the current review was to critically assess the quality of morphologic data in 14 papers that described the experimental exposure of fish to DCF. Areas of focus during this review included study design, diagnostic accuracy, magnitude of reported changes, data interpretation and presentation, and the credibility of individual reported findings. Positive attributes of some studies included robust experimental designs, accurate diagnoses, and straightforward and transparent data reporting. Issues identified in certain articles included diagnostic errors, failure to account for sampling and/or observer bias, failure to evaluate findings according to sex, exaggeration of lesion severity, interstudy inconsistencies, unexplained phenomena, and incomplete or ambiguous data presentation. It is hoped that the outcome of this review will be of value for personnel involved in regulatory decision-making.

Human Symmetry Uncertainty Detected by a Self-Organizing Neural Network Map

Symmetry in biological and physical systems is a product of self-organization driven by evolutionary processes, or mechanical systems under constraints. Symmetry-based feature extraction or representation by neural networks may unravel the most informative contents in large image databases. Despite significant achievements of artificial intelligence in recognition and classification of regular patterns, the problem of uncertainty remains a major challenge in ambiguous data. In this study, we present an artificial neural network that detects symmetry uncertainty states in human observers. To this end, we exploit a neural network metric in the output of a biologically inspired Self-Organizing Map Quantization Error (SOM-QE). Shape pairs with perfect geometry mirror symmetry but a non-homogenous appearance, caused by local variations in hue, saturation, or lightness within and/or across the shapes in a given pair produce, as shown here, a longer choice response time (RT) for “yes” responses relative to symmetry. These data are consistently mirrored by the variations in the SOM-QE from unsupervised neural network analysis of the same stimulus images. The neural network metric is thus capable of detecting and scaling human symmetry uncertainty in response to patterns. Such capacity is tightly linked to the metric’s proven selectivity to local contrast and color variations in large and highly complex image data.

Models and Methods for Discovery of Local Area Network Topology with Incomplete Data

A lot of network management tasks require a description of the logical and physical computer network topology. Obtaining such a description in an automatic way is complicated due to the possibility of incompleteness and incorrectness of the initial data on the network structure. This article provides a study on the properties of incomplete initial data on network device connectivity on the link layer. Methods for generalized handling of the heterogeneous input data on the link layer are included. We describe models and methods for deriving a missing part of the data, as well as the condition in which it is possible to get a single correct network topology description. The article includes algorithms for building a link layer topology description from incomplete data when this data is possible to fulfill up to the required level. Also, we provide methods for detecting and resolving an ambiguity in the data and methods for improving incorrect initial data. The tests and evaluations provided in the article demonstrate the applicability and effectiveness of the build methods for discovering various heterogeneous real-life networks. Additionally, we show the advantages of the provided methods over the previous analogs: our methods are able to derive up to 99% data on link layer connectivity in polynomial time; able to provide a correct solution from an ambiguous data.

Models and methods for local area network topology discovery with incomplete data

A lot of network management tasks require a description of the logical and physical computer network topology. Obtaining such a description in an automatic way is complicated due to the possibility of incompleteness and incorrectness of the initial data on the network structure. This article provides a study on the properties of incomplete initial data on network device connectivity on the link layer. Methods for generalized handling of the heterogeneous input data on the link layer are included. We describe models and methods for deriving a missing part of the data, as well as the condition in which it is possible to get a single correct network topology description. The article includes algorithms for building a link layer topology description from incomplete data when this data is possible to fulfill up to the required level. Also, we provide methods for detecting and resolving an ambiguity in the data and methods for improving incorrect initial data. Tests and evaluations provided in the article demonstrate the applicability and effectiveness of the build methods for discovering various heterogeneous real-life networks. Additionally, we show advantages of the provided methods over the previous analogs: our methods are able to derive up to 99\% data on link layer connectivity in polynomial time; able to provide a correct solution from an ambiguous data.

Human Symmetry Uncertainty Detected by a Self-Organizing Neural Network Map

Symmetry in biological and physical systems is a product of self-organization driven by evolutionary processes, or mechanical systems under constraints. Symmetry-based feature extraction or representation by neural networks may unravel the most informative contents in large image databases. Despite significant achievements of artificial intelligence in recognition and classification of regular patterns, the problem of uncertainty remains a major challenge in ambiguous data. In this study, we present an artificial neural network that detects symmetry uncertainty states in human observers. To this end, we exploit a neural network metric in the output of a biologically inspired Self-Organizing Map, the Quantization Error (SOM-QE). Shape pairs with perfect geometric mirror symmetry but a non-homogenous appearance, caused by local variations in hue, saturation, or lightness within and/or across the shapes in a given pair produce, as shown here, longer choice RT for ‘yes’ responses relative to symmetry. These data are consistently mirrored by the variations in the SOM-QE from unsupervised neural network analysis of the same stimulus images. The neural network metric is thus capable of detecting and scaling human symmetry uncertainty in response to patterns. Such capacity is tightly linked to the metric’s proven selectivity to local contrast and color variations in large and highly complex image data.