## About

9

Publications

4,155

Reads

**How we measure 'reads'**

A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more

18

Citations

Citations since 2017

Introduction

**Skills and Expertise**

Additional affiliations

May 2006 - May 2007

**Itella TGM**

Position

- Researcher

Description

- Machine Learning. Analytical CRM, Customer scoring and classification using census and other data. Mathematical prediction models for development of cities and customership. Residential zones classification. MATLAB, Excel, C++, databases.

Education

October 1999 - May 2006

## Publications

Publications (9)

We introduce in this paper methods for finding mutually corresponding dependent components from two different but related data sets in an unsupervised (blind) manner. The basic idea is to generalize cross-correlation analysis for taking into account higher-order statistics. We propose independent component analysis (ICA) type extensions for the sin...

Calculating partial derivates dx/dz and dy/dz of complex valued function f(z) = x(z) + i*y(z) is non-trivial calculation because scaling factor is not 1 but 1/2. I show two ways to calculate these derivates, first one uses Wirtinger calculus and other one Cauchy-Riemann equations. These results are well-cited in the literature and this paper works...

The multilayer neural network architecture has seen relatively little number-theoretic research. However, improving fundamental computing capacity by extending to the more extensive set of possible numbers and functions could improve machine learning results. Typically, networks often use real and sometimes complex numbers, and extending numbers to...

In recent years, the ideas of the political right, where, for example, selfishness is good for society and that there should not be many taxes to have social safety nets, have gained more ground. However, there are mathematical results that show that thinking, in which only the strongest should survive, may be partly wrong. In this article, we desc...

Delineation of drainage basins from a digital elevation model (DEM) has become a standard operation in a number of terrain
analysis software packages, but limitations of the conventionally used techniques have become apparent. Firstly, the delineation
methods make assumption of error-free data, which is an unreachable utopia even with modern sensor...

Delineation of drainage basins is a popular terrain analysis method for digital elevation model (DEM) data. Currently, a deterministic delineation method is available in a number of terrain analysis software applications. The method uses elevation data for defining flow directions for each elevation point of a DEM and then follows flow paths from a...

Automatic delineation of drainage basins from digital elevation models (DEMs) is a well established technique used in terrain analysis. The conventional methodological framework was first developed in the 1980s, after which time complexities and memory requirements of the algorithms for N-cell DEMs have been improved to the point where they are opt...

In this paper, we introduce some methods for finding mutually corresponding dependent components from two different but related data sets in an unsupervised (blind) manner. The basic idea is to generalize cross-correlation analysis by taking into account higher-order statistics. We propose independent component analysis (ICA) type extensions for th...

## Questions

Questions (10)

Hello

I’m Scientist but psychiatric hospital claims I must be in hospital and use many injection medicines!

They used strong words when I talked about university level physics to them and don’t want to think I’m Scientist and Analyst.

They might soon destroy me totally and kill me.

I’m trying to fit some kind of causal model to continuous value data by solving differential equations probabilistically (machine learning).

Currently I’m solving complex-valued vector quadratic differential equation so there are more cross correlations between variables.

dx(t)/dt = diag(Ax(t)x(t)^h) + Bx(t) + c + f(t)

or just

dx(t)/dt = diag(Ax(t)x(t)^h) + Bx(t) + c

diag() takes diagonal of the square matrix.

But my diff. eq. math is rusty because I have studied differential equations 20 years ago. I solved the equation in 1-dimensional case but would need help for vector valued x(t).

Would someone point me to appropriate material?

EDIT: I did edit the question to be a bit more clear to read.

I’m looking for a computer controlled (bluetooth/wireless) TENS/EMS unit that would have SDK to control electronic stimulation strenght and timing. I would use this for research to test dynamic stimulation techniques based on other sensor values.

All the TENS/EMS units I find online don’t have access to SDK and one must use predefined programs.

Do you know wireless TENS/EMS units that you could control from computer?

I'm looking for a cheap server with 2 TB or more RAM memory to test a novel machine learning algorithm that gives auspicious results when running with 64 GB of RAM. By interpolating current results, it seems that the machine learning algorithm would start to give good results after 1 TB or more memory.

If the algorithm works, the plan is to publish the results in a scientific journal (low impact factor thought to get the paper accepted) and develop C++ implementation in my Dinrhiw2 machine learning library (open source). Amazon AWS sells such a server at 5 USD/hour level meaning approx 840 USD/week (too much) (one month of computational time should be enough).

I’m defining a number system where numbers form a polynomial ring with cyclic convolution as multiplication.

My DSP math is a bit rusty so I’m asking when does inverse circular convolution exist? You can easily calculate it using FFT but I’m uncertain when does it exist? I would like to have a full number system where each number only has a single well defined inverse. Another part of my problem is derivation. Let c be number in my number system C[X] where coefficients are complex numbers. Linear functions can be derivated easily but I’m struggling to minimize mean squared error (i = 0..degree(C[X]), s_i(c) selects i:th coefficient of the number, s_i(x): C[x]->C):

error(W) = Exy{sum(i)(0.5|s_i(Wx - y)|^2)}

I can solve problem in case of complex numbers W E C but not in case of W E C[X] where multiplication is circular convolution. In practice my linear neural network code diverges to infinity when I try to minimize squared error.

Pointing any online materials that would help would be great.

I am implementing neural network code using C++ so I don’t have access to automatic differentation.

I can calculate gradients/backpropagation for (recurrent) feedforward neural networks easily but have trouble calculating same formulas for residual neural networks (ResNet).

Would somebody point me to online resource which would show step-by-step calculation of gradient/backprop for residual networks (ResNet)?

Here is a link to pdf that contains my calculations for neural network gradients (attempt for skip layers too):

Where I could find a good information about crypted computations (in servers)?

For example, in linear algebra you can send crypted parameters B’ and x’ to a server for computing y’ = B’*x’ where original parameters are crypted using matrix C, x’ = C*x and B’ = C*A*inv(C). So after receiving the crypted result y’ from the server, the client can then get results by computing y=inv(C)*y’.

In relation to this I’m interested in properties of function space of invertible functions F = {f | f^-1 exists}.

Mathematically speaking, optimal groups should always win group of individuals behaving optimally. However, in practice game theory is often used which assumes individuals just try to maximize their own good as well as possible.

I have only little information about game theory but recently tried to calculate information theoretic transfer rates (evolutionary adaption speed) of optimal groups when individuals compete against each other and when fitness for having children is determined through some other means.

Could someone point me to *GOOD* books/research material related to optimal groups and game theory? I would not like to waste my time reading some random on game theory.

Does any one know high-quality information about nervous system attack, control and manipulation, reading methods that intelligence services and others have developed or are actively developing (www.mindcontrol.se but it seems to be a hoax)? [potentially very dangerous]

PS. I do not feel very well by myself sometime since I have been interested in these topics.