🔬📒 Publications

• Why would you sort when you know where things approximately belong?, Authorea, March 2025
• C++ std::sort implementation is slow [O(n log n)]. With few lines of code, you can make it faster [O(n log log n) - confirmed theoretically and empirically in the publication]. Works for Gaussian, uniform, skewed and many other distributions [not included in the publication].
• Optimal blending of multiple independent prediction models, Frontiers, February 2023
• Did you ever wonder how many features you need in statistics to get to 100% truth? Turns out, the answer is infinity. Statistically, nothing is 100% truth. [Axioms in mathematics are, but mathematics is disconnected from our universe and lives in its own. In our universe, mathematics is used to approximate reality.]. Our universe can only be understood in infinity, which means never. Beauty of life - there will always be something new to invent.
• Stability of linear second-order time-varying differential equations via contractive polygons, Elsevier, April 2022
• Stability for differential equations with constant coefficients is well known. But what if coefficients are time varying? Levin came up with stability condition for such systems, but while his condition is sufficient, it can be extended in two dimensions and most likely also in higher dimensions.
• Linear regression on a set of selected templates from a pool of randomly generated templates, Elsevier, December 2021
• Linear regression is powerful tool and can often match deep neural networks performance. How do you improve machine learning algorithms? Turns out you should go wider.
• Word Replaceability Through Word Vectors, Springer, April 2020
• Imagine someone gives you a lot of text in foreign language you don't understand. Turns out, you can find synonyms in that text without understanding a single word. [Getting ready for aliens...] Word vectors come handy.
• Clustering for Binary Featured Datasets, Springer, October 2018
• What about clustering which automatically finds number of clusters? Extended publication with deeper analysis.
• Powered Outer Probabilistic Clustering, IAENG, October 2017
• What about clustering which automatically finds number of clusters? Usually in optimization, local minimum is not a good thing. In this publication, we use it to separate clusters. 
• Optimal Control Mesh, IAENG, October 2012
• Finding optimal control for non linear differential equations is not an easy task. This algorithm is way faster than reinforcement learning as algorithm spreads the solution backwards in time directly without trying many, many solutions like reinforcement learning does.
• Subdivision algorithm for optimal control, Wiley, March 2012
• Slower than Optimal Control Mesh, but sometimes you might want to double check results.
• Kneser–Ney Smoothing With a Correcting Transformation for Small Data Sets, IEEE, August 2007
• Challenge accepted. Should work also on higher n-grams, but for larger datasets.
• Approximation of box dimension of attractors using the subdivision algorithm, Taylor & Francis, January 2007
• Our world is chaotic system. It's good to know it and understand it.
• On stability of linear time-varying second-order differential equations, Jstor - Brown University, March 2006
• Stability for differential equations with constant coefficients is well known. But what if coefficients are time varying?
• MPC control using AR-Volterra models, IEEE, June 2003
• Volterra models can describe many nonlinear systems and model predictive control can find directly control for Volterra models.