arXiv subjects

Alix Petit

Publications and source records attributed to Alix Petit.

6 recordsLinked to original sources

arXiv2026

Set Shaping Theory Applied to Universal Coding

Universal coders process individual sequences without assuming that the source distribution is known. In this setting, uniformly generated sequences represent the most difficult test case: the source simulates pure randomness, contains no exploitable bias, and forces a frequency-estimating universal coder to infer the empirical composition entirely from the sequence itself. This paper reports that a Set Shaping Theory (SST) transformation systematically reduces the average universal coding length of uniformly generated sequences below the Krichevsky-Trofimov baseline N H_0(s) + R_KT(s). The transformation maps each input sequence s in A^N into an expanded sequence f(s) in A^(N+1), while storing the transformation index in the additional symbol in order to preserve reversibility. The comparison evaluates the exact Krichevsky-Trofimov baseline N H_0(s) + R_KT(s) against the shaped score (N+1)H_0(f(s)) + R_KT(f(s)), where H_0 is computed from the empirical frequencies of the individual sequence. A single unified transformation also yields reductions across distinct compression architectures, including adaptive arithmetic coding, enumerative coding, LZ78, adaptive Huffman coding, and adaptive ANS. These results support the interpretation of SST as a representation-level preprocessing layer that can structurally improve existing universal coders without requiring internal modifications to their coding mechanisms. All results reported in the article can be reproduced with the simulator available at https://sst-simulator.github.io/Set-Shaping-Theory-Simulator/.

cs.IT↗

arXiv2026

Using Set Shaping Theory to Trade RAM Accesses for CPU Computation

This paper studies Set Shaping Theory (SST) in a database-index setting under a revised interpretation: SST is not treated as a competing hashing method, but as a structural pre processing layer that can be applied before an existing indexing algorithm. The experimental question is therefore whether a method improves when it is used with SST rather than with out it. The study compares linear probing, double hashing, quadratic probing, and Robin Hood hashing against their corresponding SST-augmented variants for shaping orders K = 2,4,8. Beyond mean time, the benchmark reports mean successful probes, 95th and 99th percentile probes, collisions per stored record, and maxi mum cluster length. Experiments cover load factors from 0.75 to 0.95, database sizes from M =5000 to M =500000, query multipliers up to 200 lookups per stored record, and both uniform and hotspot query distributions. The results highlight two fundamental advantages. First, SST reduces the number of RAM accesses required during retrieval. By prevent ing clusters and long probe chains from forming at insertion time, the lookup phase requires fewer memory jumps, lower probe counts, and reduced tail latency. Second, the method introduces a new way of thinking about data storage: the data are not treated as fixed objects that must be placed passively into a table, but as reversible representations that can be struc turally adapted before being written. A small metadata tag records which transformation was selected, allowing the original key to remain recoverable and the lookup process to remain deterministic.This article is connected to the Set Shaping Theory simulator project, available online at https://sst-simulator.github.io/Set-Shaping-Theory-Simulator/ where it is possible to simulate part of the results presented in the article.

cs.IT↗

arXiv2025

Set Shaping Theory and the Foundations of Redundancy-Free Testable Codes

To render a sequence testable, namely capable of identifying and detecting errors, it is necessary to apply a transformation that increases its length by introducing statistical dependence among symbols, as commonly exemplified by the addition of parity bits. However, since the decoder does not have prior knowledge of the original symbols, it must treat the artificially introduced symbols as if they were independent. Consequently, these additional symbols must be transmitted, even though their conditional probability, under ideal and error free conditions, would be zero. This sequence extension implies that not all symbol combinations of the new length are practically realizable: if an error modifies a sequence, making it inadmissible such an error becomes detectable. Recent developments in Set Shaping Theory have revealed a surprising result: it is always possible to transform a sequence into a longer version by carefully selecting which longer sequences are allowed, in such a way that the overall set of sequences becomes more structured and less complex than the original. This means that even though the sequence is extended and dependencies are introduced between symbols, the total amount of information contained in the new set does not increase proportionally on the contrary, it can be slightly reduced. In other words, one can construct a new set of longer sequences where each one corresponds uniquely to an original sequence, but the entire set is designed in such a way that it can be treated as if the symbols were independent, making encoding simpler. This allows sequence to become testable capable of detecting errors without adding visible redundancy or increasing the informational content.

cs.IT↗

arXiv2023

Overcoming the compression limit of the individualsequence (zero order empirical entropy) using the Set Shaping Theory

Given the importance of the claim, we want to start by exposing the following consideration: this claim comes out more than a year after the article "Practical applications of Set Shaping Theory in Huffman coding" which reports the program that carried out an experiment of data compression in which the coding limit NH0(S) of a single sequence was questioned. We waited so long because, before making a claim of this type, we wanted to be sure of the consistency of the result. All this time the program has always been public; anyone could download it, modify it and independently obtain the reported results. In this period there have been many information theory experts who have tested the program and agreed to help us, we thank these people for the time dedicated to us and their precious advice. Given a sequence S of random variables i.i.d. with symbols belonging to an alphabet A; the parameter NH0(S) (the zero-order empirical entropy multiplied by the length of the sequence) is considered the average coding limit of the symbols of the sequence S through a uniquely decipherable and instantaneous code. Our experiment that calls into question this limit is the following: a sequence S is generated in a random and uniform way, the value NH0(S) is calculated, the sequence S is transformed into a new sequence f(S), longer but with the symbols belonging to the same alphabet, finally we code f(S) using Huffman coding. By generating a statistically significant number of sequences we obtain that the average value of the length of the encoded sequence f(S) is less than the average value of NH0(S). In this way, a result is obtained which is incompatible with the meaning given to NH0(S).

cs.IT↗

arXiv2023

Reviewed of the compression limit of an individual sequence using the Set Shaping Theory

Abstract: In this article, we will analyze in detail the coding limit of an individual sequence by introducing the latest developments brought by the Set Shaping Theory. This new theory made us realize that there is a huge difference between source entropy and zero order empirical entropy. Understanding the differences between these two variables allows us to take an important step forward in the study of the compression limit of an individual sequence, which we know is not calculable.

cs.IT↗

arXiv2022

Practical applications of Set Shaping Theory in Huffman coding

One of the biggest criticisms of the Set Shaping Theory is the lack of a practical application. This is due to the difficulty of its application. In fact, to apply this technique from an experimental point of view we must use a table that defines the correspondences between two sets. However, this approach is not usable in practice, because the table has A^N elements, with A number of symbols and N length of the message to be encoded. Consequently, these tables can be implemented in a program only when A and N have a low value. Unfortunately, in these cases, there are no compression algorithms with such efficiency as to detect the improvement introduced by this method. In this article, we use a function capable of performing the transform without using the correspondence table; this allows us to apply this theory to a wide range of values of A and N. The results obtained confirm the theoretical predictions.

cs.IT↗