An automatic cryptanalysis of transposition ciphers using compression

Abstract

Automatically recognising valid decryptions as a result of ciphertext only cryptanalysis of simple ciphers is not an easy issue and still considered as a taxing problem. In this paper, we present a new universal compression-based approach to the automatic cryptanalysis of transposition ciphers. In particular, we show how a Prediction by Partial Matching (PPM) compression model, a scheme that performs well at many language modelling tasks, can be used to automatically recognise the valid decrypt with a 100% success rate. We also show how it significantly outperforms another compression scheme, Gzip. In this paper, we propose a full mechanism for the automatic cryptanalysis of transposition ciphers which also automatically adds spaces to decrypted texts, again using a compression-based approach, in order to achieve readability.