Application of Divisors on a Hyperelliptic Curve in Python

The paper studies hyperelliptic curves of the genus g > 1, divisors on them and their applications in Python programming language. The basic necessary definitions and known properties of hyperelliptic curves are demonstrated, as well as the notion of polynomial function, its representation in unique form, also the notion of rational function, norm, degree and conjugate to a polynomial are presented. These facts are needed to calculate the order of points of desirable functions, and thus to quickly and efficiently calculate divisors. The definition of a divisor on a hyperelliptic curve is shown, and the main known properties of a divisor are given. There are also an example of calculating a divisor of a polynomial function, reduced and semi-reduced divisors are described, theorem of the existence of such a not unique semi-reduced divisor, and theorem of the existence of a unique reduced divisor, which is equivalent to the initial one, are proved. In particular, a semi-reduced divisor can be represented as an GCD of divisors of two polynomial functions. It is also demonstrated that each reduced divisor can be represented in unique form by pair of polynomials [a(x), b(x)], which is called Mumford representation, and several examples of its representation calculation are given. There are shown Cantor’s algorithms for calculating the sum of two divisors: its compositional part, by means of which a not unique semi-reduced divisor is formed, and the reduction part, which gives us a unique reduced divisor. In particular, special case of the compositional part of Cantor’s algorithm, doubling of the divisor, is described: it significantly reduces algorithm time complexity. Also the correctness of the algorithms are proved, examples of applications are given. The main result of the work is the implementation of the divisor calculation of a polynomial function, its Mumford representation, and Cantor’s algorithm in Python programming language. Thus, the aim of the work is to demonstrate the possibility of e↵ective use of described algorithms for further work with divisors on the hyperelliptic curve, including the development of cryptosystem, digital signature based on hyperelliptic curves, attacks on such cryptosystems.

Randomized Primality Tests

Randomization and probabilistic approach in the algorithms development occupy prominent place.Due to limited computing resources and complexity many tasks in some cases it’s impossible to obtain accurate results or it’s too costly, so the results may contain some uncertainty. There are also cases when the indeterminacy of the algorithm is its advantage, for example in cryptography problems, or a useful characteristic: in simulations of processes containing undefined parameters.In this paper, we consider the basic concepts and statements concerning randomized algorithms for checking numbers for simplicity, we present the necessary theorems.

Zero Forcing Number of Some Families of Graphs

The work is devoted to the study of the zero forcing number of some families of graphs. The concept of zero forcing is a relatively new research topic in discrete mathematics, which already has some practical applications, in particular, is used in studies of the minimum rank of the matrices of adjacent graphs. The zero forcing process is an example of the spreading process on graphs. Such processes are interesting not only in terms of mathematical and computer research, but also interesting and are used to model technical or social processes in other areas: statistical mechanics, physics, analysis of social networks, and so on. Let the vertices of the graph G be considered white, except for a certain set of S black vertices. We will repaint the vertices of the graph from white to black, using a certain rule.Colour change rule: A white vertex turns black if it is the only white vertex adjacent to the black vertex.[5] The zero forcing number Z(G) of the graph G is the minimum cardinality of the set of black vertices S required to convert all vertices of the graph G to black in a finite number of steps using the ”colour change rule”.It is known [10] that for any graph G, its zero forcing number cannot be less than the minimum degree of its vertices. Such and other already known facts became the basis for finding the zero forcing number for two given below families of graphs:A gear graph, denoted W2,n is a graph obtained by inserting an extra vertex between each pair of adjacent vertices on the perimeter of a wheel graph Wn. Thus, W2,n has 2n + 1 vertices and 3n edges.A prism graph, denoted Yn, or in general case Ym,n, and sometimes also called a circular ladder graph, is a graph corresponding to the skeleton of an n-prism.A wheel graph, denoted Wn is a graph formed by connecting a single universal vertex to all vertices of a cycle of length n.In this article some known results are reviewed, there is also a definition, proof and some examples of the zero forcing number and the zero forcing process of gear graphs and prism graphs.

Simulating Stochastic Diffusion Processes and Processes with “Market” Time

The paper focuses on modelling, simulation techniques and numerical methods concerned stochastic processes in subject such as financial mathematics and financial engineering. The main result of this work is simulation of a stochastic process with new market active time using Monte Carlo techniques.The processes with market time is a new vision of how stock price behavior can be modeled so that the nature of the process is more real. The iterative scheme for computer modelling of this process was proposed.It includes the modeling of diffusion processes with a given marginal inverse gamma distribution. Graphs of simulation of the Ornstein-Uhlenbeck random walk for different parameters, a simulation of the diffusion process with a gamma-inverse distribution and simulation of the process with market active time are presented.To simulate stochastic processes, an iterative scheme was used: xk+1 = xk + a(xk, tk) ∆t + b(xk, tk) √ (∆t) εk,, where εk each time a new generation with a normal random number distribution.Next, the tools of programming languages for generating random numbers (evenly distributed, normally distributed) are investigated. Simulation (simulation) of stochastic diffusion processes is carried out; calculation errors and acceleration of convergence are calculated, Euler and Milstein schemes. At the next stage, diffusion processes with a given distribution function, namely with an inverse gamma distribution, were modelled. The final stage was the modelling of stock prices with a new "market" time, the growth of which is a diffusion process with inverse gamma distribution. In the proposed iterative scheme of stock prices, we use the modelling of market time gains as diffusion processes with a given marginal gamma-inverse distribution.The errors of calculations are evaluated using the Milstein scheme. The programmed model can be used to predict future values of time series and for option pricing.

Generalization of Cross-Entropy Loss Function for Image Classification

Classification task is one of the most common tasks in machine learning. This supervised learning problem consists in assigning each input to one of a finite number of discrete categories. Classification task appears naturally in numerous applications, such as medical image processing, speech recognition, maintenance systems, accident detection, autonomous driving etc.In the last decade methods of deep learning have proven to be extremely efficient in multiple machine learning problems, including classification. Whereas the neural network architecture might depend a lot on data type and restrictions posed by the nature of the problem (for example, real-time applications), the process of its training (i.e. finding model’s parameters) is almost always presented as loss function optimization problem.Cross-entropy is a loss function often used for multiclass classification problems, as it allows to achieve high accuracy results.Here we propose to use a generalized version of this loss based on Renyi divergence and entropy. We remark that in case of binary labels proposed generalization is reduced to cross-entropy, thus we work in the context of soft labels. Specifically, we consider a problem of image classification being solved by application of convolution neural networks with mixup regularizer. The latter expands the training set by taking convex combination of pairs of data samples and corresponding labels. Consequently, labels are no longer binary (corresponding to single class), but have a form of vector of probabilities. In such settings cross-entropy and proposed generalization with Renyi divergence and entropy are distinct, and their comparison makes sense.To measure effectiveness of the proposed loss function we consider image classification problem on benchmark CIFAR-10 dataset. This dataset consists of 60000 images belonging to 10 classes, where images are color and have the size of 32×32. Training set consists of 50000 images, and the test set contains 10000 images.For the convolution neural network, we follow [1] where the same classification task was studied with respect to different loss functions and consider the same neural network architecture in order to obtain comparable results.Experiments demonstrate superiority of the proposed method over cross-entropy for loss function parameter value α < 1. For parameter value α > 1 proposed method shows worse results than cross-entropy loss function. Finally, parameter value α = 1 corresponds to cross-entropy.

Mathematical Models of Reinsurance

Insurance provides financial security and protection of the independence of the insured person. Its principles are quite simple: insurance protects investments, life and property. You regularly pay a certain amount of money in exchange for a guarantee that in case of unforeseen circumstances (accident, illness, death, property damage) the insurance company will protect you in the form of financial compensation.Reinsurance, in turn, has a significant impact on ensuring the financial stability of the insurer. Because for each type of insurance there is a possibility of large and very large risks that one insurance company can not fully assume. In the case of a portfolio with very high risks, the company may limit their acceptance, or give part of the reinsurance. The choice of path depends entirely on the company’s policy and type of insurance.This paper considers the main types of reinsurance and their mathematical models. An analysis of the probability of bankruptcy and the optimal use of a particular type of reinsurance are provided.There are also some examples and main results of research on this topic. After all, today the insurance industry is actively gaining popularity both in Ukraine and around the world. Accordingly, with a lot of competition, every insurer wants to get the maximum profit with minimal e↵ort.