Research Interests

• Machine Learning, Data Mining, Statistical Inference, Bayesian Inference, Maximum Entropy Method, Prior Choice, Model Selection
• Factorization Algorithms, Riemann Hypothesis, Riemann Zeta Function
• Integer Series
• Genetic Programming Algorithms
• High-Precision Computing

Publications

Current Interests

• "Bayesian Inference Featuring Entropic Priors" in AIP proceedings volume 954 (preprint)
  A proposal for the choice of priors when the data model is known. It is argued that application of the Maximum Entropy Method strictly requires an assumption (expectation) on the data likelihood entropy. When we don't have a clue about that, the data likelihood entropy variance should be maximized.

• I also had a draft about "Entropic Priors and Classical Estimation Theory" on this page, but I withdraw it momentarily because I noticed that I was a little bit too premature about the "classical properties" of entropic priors and my proposed alpha estimate.

Other

• "Eine PAC-Agenten-API für Graphische Bedienungsoberflächen". Studienarbeit, Lineas Integral Software GmbH / Technische Universität Braunschweig, Braunschweig (Germany), July 2002
• "Enabling Access to a Circuit Generator Library via the Flexible API for Module-based Environments". Diploma Thesis, Technische Universität Braunschweig, Braunschweig (Germany), November 2002
• Tilman Neumann, Andreas Koch: "A Generic Library for Adaptive Computing Environments." Proc. 11th Int. Conf. on Field-Programmable Logic and Applications (FPL), Belfast. Published in Springer Lecture Notes in Comp. Sci. 2147 (2001), 503-512.

OEIS Contributions

• A145176
• A145177
• A145178
• A145201
• A145518 (more terms)
• A145519
• A145520 (more terms)
• A164961 (more terms)

Software

• Java generator for partitions of natural numbers and partition transforms that generate number triangles from sequences. The package depends on the apfloat library and Apache log4j.