Voynich NLP analysis
Learning and science
linguistics research
SHOW HN2025
Legend
Voynich NLP analysis
I modeled the Voynich Manuscript with SBERT to test for structure
github.com#voynich manuscript#sbert embeddings#nlp#kmeans clustering#markov transition
How HN took it
5 positive, 17 neutral or asking, 1 mixed, 5 critical, out of 28 comments about the project. Thread score 0.00.
Idea: 3 for, 4 againstExecution: 1 for, 1 againstDocs: 1 for, 0 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Interesting, nerdy deep-dive into the Voynich
From 1 comment
This is hands-down the nerdiest and coolest deep-dive into the Voynich I’ve seen.
pawanjswal on HN
Fascinating approach to an unsolved puzzle
From 1 comment
Voynich is one of my favorite unsolved puzzles. This approach looks fascinating, so thanks for sharing your work here!
thearn4 on HN
Objections
Embedding model is outdated and may not suit an unknown language
From 1 comment
A point of note is that the text embeddings model used here is paraphrase-multilingual-MiniLM-L12-v2…
minimaxir on HN
Stripping suffixes removes information and may harm results
From 2 comments
A point of note is that the text embeddings model used here is paraphrase-multilingual-MiniLM-L12-v2…
minimaxir on HNThis is very interesting. You should post a link to https://www.voynich.ninja/index.php I'm not familiar with SBERT, or with modern statistical NLP in general, but SBERT works on…
DonaldFisk on HN
Trained-on-known-languages models cannot extract meaning from an unknown script
From 1 comment
I feel like we are missing an important point...the transformer model that has been used here is trained on known languages. This means it cannot extract meaningful embeddings…
theRealEros on HN
Manuscript is likely a hoax or gibberish
From 2 comments
I strongly believe the manuscript is undecipherable in the sense thats it's all gibberish. I can't prove it, but at this point I think it's more likely than not to be hoax.
glimshe on HNtheres no need to do any of this, its fake, its a forgery
mach5 on HN
Requests
Try UMAP or t-SNE instead of PCA
From 1 comment
UMAP or TSNE would be nice, even if PCA already shows nice separation. Reference mapping each cluster to all the others would be a nice way to indicate that there's no…
tetris11 on HN
Run a control test on known or meaningless text
From 2 comments
(I know nothing about NLP) Does it make sense to check the process with a control group? E.g. if we ask a human to write something that resembles a language but isn’t, then…
thih9 on HNThis is very interesting. You should post a link to https://www.voynich.ninja/index.php I'm not familiar with SBERT, or with modern statistical NLP in general, but SBERT works on…
DonaldFisk on HN
Try newer embedding models or keep suffixes
From 2 comments
A point of note is that the text embeddings model used here is paraphrase-multilingual-MiniLM-L12-v2…
minimaxir on HNSorry if I missed it, but what about keeping the suffixes and trying to do some finetuning on the source then clustering sentences or at least pages which given the media should…
gthompson512 on HN
Questions
Was the write-up link broken
From 1 comment
The link to the write-up seems broken, can you write the correct one?
GTP on HN
Was the text processed left to right when stripping suffixes
From 1 comment
This looks very interesting - nice work! I have no background in NLP or linguistics, but I do have a question about this: > I stripped a set of recurring suffix-like endings from…
andyjohnson0 on HN
Has the manuscript already been deciphered
From 1 comment
Wasn't it already deciphered, though? https://www.researchgate.net/publication/368991190_The_Voynich_Manuscript_Decoded
ablanton on HN
Alternatives named
Prior Voynich analyses by Zattera, Roe, Feaster and Coons
From 1 comment
The best work on Voynich has been done by Emma Smith, Coons and Patrick Feaster, about loops and QOKEDAR and CHOLDAIIN cycles. Here's a good presentation:…
veqq on HN
voynich.nu transcriptions and voynich.ninja forum
From 2 comments
The best work on Voynich has been done by Emma Smith, Coons and Patrick Feaster, about loops and QOKEDAR and CHOLDAIIN cycles. Here's a good presentation:…
veqq on HNAlthough I skimmed the methodology out of curiosity, what really drew my eye was the transcription in the repository of the manuscript. This led me down a rabbit hole leading…
andrewla on HN