kylebgorman

6 ms·

1.
Python Ellipses Considered Harmful(wellformedness.com)
2 points by kylebgorman 2y ago|2 comments
2.
by kylebgorman 4y ago
Yes, this is a discussion forum, so while you're welcome to pose ethical questions, you should also feel free to weigh in on them too ;) It's a challenging question. Is it the mere fact of someone being aroused by children that ma
3.
Re.compile Is Otiose(wellformedness.com)
1 points by kylebgorman 4y ago|0 comments
4.
by kylebgorman 4y ago
A speech recognizer is not an "AI".
5.
by kylebgorman 4y ago
The first author's an independent scholar so I don't think that's right.
6.
by kylebgorman 4y ago
I have the same set-up: SomaFM (and a Pi with an external hard disk) and three Squeezeboxes all in sync.
7.
A* shortest string decoding for non-idempotent semirings(arxiv.org)
2 points by kylebgorman 4y ago|0 comments
8.
by kylebgorman 5y ago
If I may, there's no real reason to break out ACL vs. NAACL vs. EMNLP, since they're all run by the ACL and one would be hard-pressed to say how the EMNLP community might differ from the ACL community at this point. And if you&#x2
9.
by kylebgorman 5y ago
I have been doing this for a long time, and FreeDB's metadata is way, way worse than MusicBrainz's, and has less coverage too. If you find any issues in MusicBrainz metadata you can submit a correction, which will be reviewed and
10.
by kylebgorman 5y ago
Sorry but this person isn't nearly famous enough to have been "cancelled".
11.
by kylebgorman 5y ago
Get W. Sidney Allen's slim text _Vox Graeca_. It goes over not only how Ancient Greek was pronounced but also how we know. The evidence includes explicit notes from grammarians and other writers, graffiti, borrowings, puns, and later f
12.
Finite-state text processing in Python(pynini.opengrm.org)
3 points by kylebgorman 5y ago|0 comments
13.
Bayesian statistical model comparison for natural language processing(arxiv.org)
1 points by kylebgorman 6y ago|0 comments
14.
Translating lost languages using machine learning(wellformedness.com)
2 points by kylebgorman 6y ago|0 comments
15.
by kylebgorman 8y ago
"text normalization" just refers to that sort of thing. it's a subcomponent of the "front-end" of a speech synthesizer but it can also be used for speech recognition (if your training data contains things like &quot
16.
by kylebgorman 9y ago
Mechanical reproduction is a force multiplier allowing your art to touch millions or billions. (Mozart & Michelangelo didn't know this was coming, but I'm a consequentialist, so I give them credit for it.) Far smaller is the a
17.
by kylebgorman 9y ago
Author should have a native English speaker read this for fluency...
18.
by kylebgorman 9y ago
Activist investors who want changes in business model or cost are extremely common. Just about every publicly traded company has to deal with one at some point.
19.
What to do about the academic brain drain(wellformedness.com)
1 points by kylebgorman 9y ago|0 comments
20.
by kylebgorman 9y ago
I am building libraries for predicting which pronunciation of a homograph to use for in a text-to-speech application. E.g., when we go to synthesize "bass", should we use the pronunciation that rhymes with "mace" or &quo
21.
by kylebgorman 9y ago
He'd led a violent (albeit somewhat pathetic, and suppressed) attempt at a coup---that's why he was in jail. Rough comparison points in recent US history would be the domestic terrorists like McVeigh and Kaczynski, I suppose.
22.
by kylebgorman 10y ago
Unless you know something I don't, the SOTA in ASR is a hybrid system that still has all the complexity of the traditional generative model plus NNs on top.
23.
by kylebgorman 10y ago
As a professional linguist, I think that Farmer et al. are probably right and it's almost surely not a "writing system" at all. (This is a term of art with a very specific sense.)
24.
by kylebgorman 10y ago
They have "limo licenses" but yeah, few if any of them are actual limousines in wheelbase / shape. Most are towncars or upscale SUVs.
25.
by kylebgorman 10y ago
I would not recommend NLTK (or its book) or Jurafsky & Martin, or Manning & Schuetze. All are insanely dated. Watch some Coursera lectures, check out a newer, non-academic, application-oriented text, or just build something.
26.
by kylebgorman 10y ago
This is not a question that really would be of much interest to most working linguists, who aren't really willing to make value judgments about language any more than chemists are to talk about their favorite molecules. Constructed lan
27.
by kylebgorman 10y ago
Or: alias dronezone='mplayer http://somafm.com/dronezone130.pls' alias groovesalad='mplayer http://somafm.com/groovesalad130.pls' (etc.)
28.
by kylebgorman 10y ago
Before Apple sold music, iTunes had an Internet Radio tab that listed several SomaFM stations.
29.
by kylebgorman 10y ago
That's an interesting problem. But I'm not seeing much upside for you and there's a high chance of being laid off before you get situated.
30.
by kylebgorman 10y ago
Anything that uses the standard decoding strategy (Viterbi) is quadratic in the number of "states" (i.e., tags), though, and that easily dominates. (Nowadays, many people just use greedy---linear in the number of states---decoding

More ›