| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| shorttalksjune2021 [2022/10/30 22:44] – old revision restored (2021/04/21 13:02) 139.124.146.3 | shorttalksjune2021 [2022/10/30 22:44] (current) – old revision restored (2021/04/19 10:32) 139.124.146.3 |
|---|
| |
| In Fibonacci coding we forbid '11' patterns, bringing question of generalization to more complex constraints. I will very briefly present how to practically handle such general situation, including 2D lattice (hard square problem), with motivation e.g. to increase capacity of hard disk. We will first find probability distribution of patterns in such ensemble of sequences, which can be obtained with MERW ( https://en.wikipedia.org/wiki/Maximal_entropy_random_walk ). Then we can use entropy coder - for this purpose I have introduced ANS ( https://en.wikipedia.org/wiki/Asymmetric_numeral_systems ): generalization of numeral systems to optimize for a chosen probability distribution of digits. I will also briefly present its two widely used variants: rANS with multiplication - now e.g. in JPEG XL image format, or tANS building finite state automaton for a given probability distribution - now used e.g. in Facebook Zstandard or Apple LZFSE compressors. | In Fibonacci coding we forbid '11' patterns, bringing question of generalization to more complex constraints. I will very briefly present how to practically handle such general situation, including 2D lattice (hard square problem), with motivation e.g. to increase capacity of hard disk. We will first find probability distribution of patterns in such ensemble of sequences, which can be obtained with MERW ( https://en.wikipedia.org/wiki/Maximal_entropy_random_walk ). Then we can use entropy coder - for this purpose I have introduced ANS ( https://en.wikipedia.org/wiki/Asymmetric_numeral_systems ): generalization of numeral systems to optimize for a chosen probability distribution of digits. I will also briefly present its two widely used variants: rANS with multiplication - now e.g. in JPEG XL image format, or tANS building finite state automaton for a given probability distribution - now used e.g. in Facebook Zstandard or Apple LZFSE compressors. |
| |
| ** [[https://www.researchgate.net/profile/Palak-Pandoh|Palak Pandoh]]** //Counting scattered palindromes in a finite word// | |
| |
| We investigate the scattered palindromic subwords in a finite word. We start by characterizing the words with the least number of scattered palindromic subwords. Then, we give an upper bound for the total number of palindromic subwords in a word of length $n$ in terms of Fibonacci number $F_n$ and prove that at most $F_n$ new scattered palindromic subwords can be created on the concatenation of a letter to a word of length $n − 1$. We then show that the maximum number of scattered palindromic subwords in a word depends both on the length of the word and the number of distinct letters in the word. We finally give a tight bound on the number of scattered palindromic subwords in a word of length $n$ that has at least $n/2$ distinct letters. | |
| |
| ** [[https://www.researchgate.net/profile/Bartlomiej-Pawlik|Bartłomiej Pawlik]] ** //Nonchalant procedure// | |
| |
| Recently, [[https://www.combinatorics.org/ojs/index.php/eljc/article/view/v27i1p48|Gryczuk, Kordulewski and Niewiadomski]] introduced the concept of the nonchalant procedure. In the original form, the procedure generates square-free words over the alphabet $\{1,2,3\}$ in such manner that every word (except for the initial word "$1$") is the extension of the preceding word by inserting the first letter possible (in the natural order on respective alphabet) before the shortest suffix possible (preferably of the length $0$) resulting in the new word being square-free. Aforementioned article contains the conjecture, that such a procedure never stops. | |
| |
| While the conjecture remains open, further investigation of the subject pointed out certain "flexibility" of the nonchalant procedure - it can be considered in various contextes, not necessarily focused on combinatorics on words. In this talk we will present and analyze a few variants of the procedure. | |
| |
| ** [[https://www.utu.fi/en/people/jarkko-peltomaki|Jarkko Peltomäki]] ** //Initial nonrepetitive complexity of regular episturmian words and their Diophantine exponents// | |
| |
| Regular episturmian words are episturmian words whose | |
| directive words $x_1^{a_1} x_2^{a_2} \dotsm$ have the property that $x_1 | |
| x_2 \dotsm = u^\omega$ for a word $u$ whose all letters are distinct. | |
| This subclass of episturmian words contains all Sturmian words and | |
| generalizations of the Fibonacci word. | |
| |
| The Diophantine exponent $\mathrm{dio}(x)$ of an infinite word $x$ is | |
| the supremum $\rho$ of real numbers for which there exist arbitrarily | |
| long prefixes of $x$ of the form $UV^e$ such that $|UV^e|/|UV| \geq | |
| \rho$. This concept is especially interesting since $\mathrm{dio}(x)$ | |
| gives a lower bound for the irrationality exponent of the real number | |
| having $x$ as a fractional part. | |
| |
| In the talk, I will discuss how the Diophantine exponent of an infinite | |
| word can be found by studying its initial nonrepetitive complexity | |
| function. Then I will explain how to determine the initial nonrepetitive | |
| complexity of regular episturmian words and describe what can then be | |
| inferred of their Diophantine exponents. The main corollaries of the | |
| results are as follows. Under certain mild conditions, a real number | |
| having a regular episturmian word as a fractional part has irrationality | |
| exponent $> 2$. Moreover, such a number has irrationality exponent | |
| $\infty$ (i.e., it is a Liouville number) if and only if the | |
| corresponding directive word has unbounded partial quotients. | |
| |