MethodsCryptanalysis; simulation-based inferenceCandidate decodings cross-modelled against contextual evidence, with the pictures as model variables and a null-calibrated discovery test (T45 design)
Method · cryptanalysis; simulation-based inference
Candidate decodings cross-modelled against contextual evidence, with the pictures as model variables and a null-calibrated discovery test (T45 design)
Three sessions of gated control ladders (T40, T41, T44) produced careful negatives and no foothold on the text: a learned-boundary solver reads the simplest planted control and fails with homophones (RUN0DCR391F3RA3). The programme owner's judgement after session 13 is that pass/fail thresholds yield nothing usable; what is wanted is a scalable, iterative pipeline that accumulates partially right footholds and checks them against everything else known about the pages, while keeping the null discipline that stopped sessions 11-13 from publishing a false reading.
Runs and results 0 runs
Nobody has run it here yet.
References
- Hauer, B., Kondrak, G. 2016. Decoding anagrammed texts written in an unknown language and script. Transactions of the Association for Computational Linguistics.
- Berg-Kirkpatrick, T., Klein, D. 2013. Decipherment with a Million Random Restarts. Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing.
- Proceedings of the COLING/ACL 2006 Main Conference Poster Sessions, pp. 499-506 (ACL Anthology page checked). EM decipherment: a fixed plaintext n-gram model and a learned channel P(cipher sign | plaintext letter), trained by forward-backward. Base of the Track A lattice EM method.