K. Weber

2papers

2 Papers

CRDec 21, 2021
CryptoMiniSat Switches-Optimization for Solving Cryptographic Instances

A. -M. Leventi-Peetz, O. Zendel, W. Lennartz et al.

Performing hundreds of test runs and a source-code analysis, we empirically identified improved parameter configurations for the CryptoMiniSat (CMS) 5 for solving cryptographic CNF instances originating from algebraic known-plaintext attacks on 3 rounds encryption of the Small AES-64 model cipher SR$(3, 4, 4, 4)$. We finally became able to reconstruct 64-bit long keys in under an hour real time which, to our knowledge, has never been achieved so far. Especially, not without any assumptions or previous knowledge of key-bits (for instance in the form of side-channels, as in \cite{Mohamed2012algebraicSCA}). A statistical analysis of the non-deterministic solver runtimes was carried out and command line parameter combinations were defined to yield best runtimes which ranged from under an hour to a few hours in median at the beginning. We proceeded using an Automatic Algorithm Configuration (AAC) tool to systematically extend the search for even better solver configurations with success to deliver even shorter solving times. In this work we elaborate on the systematics we followed to reach our results in a traceable and reproducible way. The ultimate focus of our investigations is to find out if CMS, when appropriately tuned, is indeed capable to attack even bigger and harder problems than the here solved ones. For the domain of cryptographic research, the duration of the solving time plays an inferior role as compared to the practical feasibility of finding a solution to the problem. The perspective scalability of the here presented results is the object of further investigations.

LGDec 20, 2021
Scope and Sense of Explainability for AI-Systems

A. -M. Leventi-Peetz, T. Östreich, W. Lennartz et al.

Certain aspects of the explainability of AI systems will be critically discussed. This especially with focus on the feasibility of the task of making every AI system explainable. Emphasis will be given to difficulties related to the explainability of highly complex and efficient AI systems which deliver decisions whose explanation defies classical logical schemes of cause and effect. AI systems have provably delivered unintelligible solutions which in retrospect were characterized as ingenious (for example move 37 of the game 2 of AlphaGo). It will be elaborated on arguments supporting the notion that if AI-solutions were to be discarded in advance because of their not being thoroughly comprehensible, a great deal of the potentiality of intelligent systems would be wasted.