rspeer

99 ms·

961.
by rspeer 10y ago
Just because you haven't found a use for it doesn't mean it's useless. The stdlib datetime class is terrible and desperately needs to be wrapped. Arrow is a good wrapper. I don't know what you're on about with count
962.
by rspeer 10y ago
My company, Luminoso, uses Conceptnet Numberbatch as one component for interpreting the meaning of text. When the benchmarks went up, the understandability of its results did too. You get better search results, better topics, clearer visual
963.
by rspeer 10y ago
Ensemble methods have been around for decades because they are a good idea. Watson -- the Jeopardy bot, not the marketing brand -- was an ensemble method. My impression is that people avoid ensembles because they are fighting to give the im
964.
by rspeer 10y ago
There kind of are ConceptNet embeddings. They weren't called "embeddings" when they were released, because it was 2008 [1]. Here's an example: http://conceptnet5.media.mit.edu/data/5.4/assoc&#x2
965.
by rspeer 10y ago
Your claim of "I could do that" rings false. You didn't do it. You think that you could have done it because I wrote a clear enough paper telling you how. And even if you were curious and determined enough to do it, the thing
966.
by rspeer 10y ago
It's tricky because we don't have good ground truth on what different word senses there are. (WordNet is not the final answer, especially as it separates every metaphorical use of the same word into its own sense.) My experience i
967.
by rspeer 10y ago
You can retrain your distributional knowledge and keep your lexical knowledge. Moving to a new domain shouldn't mean you have to forget everything about what words mean and hope you manage to learn it again. The whole idea of Numberb
968.
by rspeer 10y ago
That said: LexVec gives quite good results on word-relatedness for using only distributional knowledge, and only from Wikipedia at that. Adding ConceptNet might give something that is more likely to be state-of-the-art.
969.
by rspeer 10y ago
As pre-built word vectors go, Conceptnet Numberbatch [1], introduced less flippantly as the ConceptNet Vector Ensemble [2], already outperforms this on all the measures evaluated in its paper: Rare Words, MEN-3000, and WordSim-353. This fac
970.
by rspeer 10y ago
Except it's MySQL, so by the time you read it out it says: id first last birth_year 1 Blaise Pascal 1623 2 Gottfried Leibniz 1646 3 Emmy Noether 1
971.
by rspeer 10y ago
In what cases would "e.g." be better than "for example", aside from when you're desperately short on space? Most of the cases I see "e.g." and "i.e." -- with or without the dots -- it means &quot
972.
by rspeer 10y ago
Google has tech support?!
973.
by rspeer 10y ago
You don't need any training data to do better than Porter2, just lexical data. If you're content to map the same surface form to the same stem regardless of context, as Porter2 would, what you need is a dictionary mapping words to
974.
by rspeer 10y ago
They used an equirectangular projection, because rectangles. But that means that some of these rotated rectangles are more like parallelograms when you look at them in another way. Mercator actually would have been a good choice because it
975.
by rspeer 10y ago
To be clear, Bosnia and Herzegovina is one country.
976.
by rspeer 10y ago
Why in the world would you smugly jump to the conclusion that he's ordering sugary drinks? He just said coffee . You can get black coffee at Starbucks.
977.
by rspeer 10y ago
GloVe might be what you're looking for: http://nlp.stanford.edu/projects/glove/
978.
by rspeer 10y ago
This is a normalization that I have used and seen other people use. I don't think it's a foregone conclusion that PMI is better for every task.
979.
by rspeer 10y ago
It definitely didn't have a "catch them all" mechanic, and that alone lets Pokémon Go have much more of a feeling of personal progress. In Ingress, everything is about taking over portals for your team. You get items with wei
980.
by rspeer 10y ago
That's standard word2vec, not an RNN.
981.
by rspeer 10y ago
It's like you're recommending a restaurant by saying that its food is "edible". All in all, this is a bad recommendation. Julia is version 0.4. Its syntax changes with every version. It is not going to be stable soon. It
982.
by rspeer 10y ago
And if you don't have time to switch from Python 2 to 3, you definitely don't have time to switch to Go.
983.
by rspeer 10y ago
The thing about a language being "a dialect with an army" is a bit tongue in cheek. Of course we'd like more consistent ways to distinguish languages that don't depend on politics. When there's a continuum (like the
984.
by rspeer 10y ago
And that's the original word. This is a word that's changing pretty rapidly because nobody's pushing back against it. No one shouts "you're stupid and wrong" at people who use the word "flammable" i
985.
by rspeer 10y ago
If the problem is that overseas investors are buying the houses, sure, that's a problem that decreases the amount of living space. But how would restrictive zoning fix that problem? And I still don't think the flipping is relevant
986.
by rspeer 10y ago
> Apartments and other attached units are now so expensive, and prices rising so steadily, that the logical thing to do is let them sit empty. I highly doubt this is going on over any long period of time. If there's a bubble like th
987.
by rspeer 10y ago
When it comes to finding a decent terminal emulator for Windows: mintty is the terminal you get when you use Git Bash, right? Because in my experience it's better than ConEmu. ConEmu makes a staggering number of off-by-one errors in pu
988.
by rspeer 10y ago
I'm not sure which distinction exactly you're drawing between free software and open source, but most open-source licenses don't aim to restrict how people use the software, for many reasons. One reason is that it's ju
989.
by rspeer 10y ago
FWIW, I should mention that I have tried one DB that seems to qualify as a graph database, in that it actually does seem to import and traverse edges efficiently, which is Blazegraph. You tend to have to search past several other things c
990.
by rspeer 10y ago
I'm not asking you to explain to me what a graph is. I've found that it's easier and faster to import and traverse edges in SQLite or PostgreSQL than in some of the systems you listed. I haven't tried RethinkDB but it so

More ›