Browsing by Author "Zajac, Pavol"

Now showing 1 - 4 of 4

listelement.badge.access-status Open Access ,
Can Artificial Intelligence solve the mysterious anagram from the church of the Poor Clares in Bratislava?
(Tartu University Library, 2024) Antal, Eugen; Zajac, Pavol; Waldispühl, Michelle; Megyesi, Beáta
A mysterious anagram was found in the Church of the Poor Clares in Bratislava, but as far as we know it has never been successfully solved. The anagram contains81 symbols, including specific diphthongsAE, CH, and GY. Unlike other anagramstypical of that age, the symbolsare not ordered alphabetically. We suspectthat a specific order of symbols is relatedto the original order of symbols in theplain text. Even with the suspected orderof letters, the number of possible plain textcandidates is too high to obtain the originaltext with standard methods. We examinealternative scoring methods based onmodern AI text similarity to improve thequality of the candidate plain text candidates.
listelement.badge.access-status Open Access ,
HCPortal: Ten Years of Development
(Tartu University Library, 2026-06-22) Antal, Eugen; Zajac, Pavol; Desenclos, Camille; Pierrot, Cécile
HCPortal is an online portal focusing on historical cryptology. It has now been in active development for ten years. During the development, many new changes and additional modules were added. We present the current state of the portal, with a focus on an overview of the main modules and their user interface specifics.
listelement.badge.access-status Open Access ,
Solving Anagrams with Integer Linear Programming
(Tartu University Library, 2025) Zajac, Pavol; Selep, Tomáš; Antal, Eugen; Antal, Eugen; Marák, Pavol
"Given some sequence of letters, an anagram is formed by changing their order to create a different text. In a historical context, anagrams were popular mainly as puzzles, but they are also connected to classical transposition ciphers. To solve an anagram means to rearrange the letter sequence to a form that is acceptable as a word or sentence in some language. In this article, we formalize the anagram solving problem. We focus on anagrams based on a simplified language model based on fixed dictionaries. We study the applicability of known methods for this problem. We propose a method of anagram solving based on integer linear programming. The new method is not strictly superior to existing methods but provides new tools to tackle the problem. The new representation shows potential for integration with Word2Vec representation of words for finding potentially meaningful anagrams in natural languages."
listelement.badge.access-status Open Access ,
Solving Historical Ciphers with AI: Analysis of GPT’s Capability in Processing and Deciphering Cryptographic Postcards
(Tartu University Library, 2026-06-22) Antal, Eugen; Pavuk, Tomáš; Zajac, Pavol; Desenclos, Camille; Pierrot, Cécile
In this case study, we focus on the capabilities of ChatGPT for assisting in the research of historical encrypted documents. Our study is based on the dataset of historical encrypted postcards, typically with simple (substitution) ciphers. We split the task into two parts: transcription and interpretation of the postcard image, and then decipherment of the transcribed text. We summarize the current capabilities of (multimodal) AI tools in general, and specifically of GPT-5.2 / GPT-4o that were used as AI tools in our analysis. Following our cryptanalytic pipeline, we show specifically how the AI tool was capable of solving selected typical exemplars from our dataset. We show that GPT-5.2 is already capable of solving most of the required tasks with acceptable quality, and can be useful even for researchers not specifically trained in prompt engineering or artificial intelligence usage.