crypto5

9 ms·

91.
by crypto5 10y ago
Where does considerable cost come from? There is a small overhead in number of characters in your program. I think that's it.
92.
by crypto5 10y ago
> it also imposes a development cost, Very very small cost.
93.
by crypto5 10y ago
Maybe this whole article has been generated by one of their models..
94.
by crypto5 10y ago
> I think you might be pleasantly surprised by recent results using DL and LSTM for building models of natural language. What exactly those models are modelling?
95.
by crypto5 10y ago
Lol, as someone who is not in academia and math, I didn't recognize your joke while reading this for the first time.
96.
by crypto5 10y ago
Wiki says it is proprietary..
97.
by crypto5 10y ago
No built in compression. Really big deal for reducing IO and DB size.
98.
by crypto5 10y ago
" Both corpora use Clozestyle questions (Taylor, 1953), which are formulated by replacing a word or phrase in a given sentence with a placeholder token." Looks like it is more task of text search, not "reading comprehension&q
99.
by crypto5 10y ago
Yes, but whole story is about that initial contract didn't have non-disclosure clause, they planned to publish their result, and government wanted to add it afterwards and force them to sign new papers, and discussion was about that. A
100.
by crypto5 10y ago
They declared. Government agent started explicitly threatening them with multiple felonies for proof disclosure and threatening 4th guy's family if he will not sign papers.
101.
by crypto5 10y ago
Start your pet project in AI, you will gain real hands-on experience and skills, will have some portfolio and will understand is it for you or not.
102.
by crypto5 10y ago
reminds me this good movie: https://en.wikipedia.org/wiki/Travelling_Salesman_(2012_film... Guys found proof for P = NP, and government declared it as national asset and government secret.
103.
by crypto5 10y ago
Looks like just text classifier, which classifies text snippets by predefined simple intent.
104.
by crypto5 11y ago
> The other 99.9 percent of the time when things go smoothly we don't even notice... Or just don't care if it really goes smoothly.
105.
by crypto5 12y ago
I think this is one of advertised advantages of deep learning: it will find useful and unobvious features in your data corpus without much effort from your side.
106.
by crypto5 12y ago
Typical usage is an inverted index for full text search. You have search token as a key, and doc_id as columns.
107.
by crypto5 12y ago
> Postgres has no facility to make it possible to do faceting (histograms). This is a pretty big deal breaker for most kinds of eCommerce. I've seen a few attempts at this, but they are incredibly complex and slow compared to other
108.
by crypto5 12y ago
> In my experiemce, Go is a breath of fresh air. Let's see where Go will be in 20 years.
109.
by crypto5 12y ago
I think standard Java library API doesn't know what algorithms are supported on particular platform, additionally you can implement your own messagedigest.
110.
by crypto5 12y ago
> none of these new planes actually come anywhere close to fulfilling the close ground support role the A-10 offers What makes you think that way? Low-end missions can be fulfilled by drones and apaches, and high-end obviously by F-35. A
111.
by crypto5 12y ago
But for most of the typical web apps, non-blocking/reactive/etc programming is absolutely unnecessary.
112.
by crypto5 12y ago
Did they essentially fire 5k with stack ranking annually before?
113.
by crypto5 13y ago
And maybe you have a chance to provide simple example of how program can be rephrased in such way that generic will not be needed?
114.
by crypto5 13y ago
> I think one huge difference is that a channel in Go is a stream of data, in which each element can cause the consuming code to wait for new responses. In your example, the future can only return one bit of data. In java you also can do
115.
by crypto5 13y ago
> Obviously there is a massive syntactic difference between your Java example and the equivalent Go code With Java 8 and lambdas, syntax will be much cleaner, i.e. you can write something like: executor.execute(() -> "step1"
116.
by crypto5 13y ago
Sure, you can replace SingleThreadExecutor by Executors.newFixedThreadPool(4)
117.
by crypto5 13y ago
Maybe somebody will find useful and relevant my pet project: https://github.com/crypto5/wikivector . It uses machine learning and wikipedia data as training set, supports 10 languages, and completely open source.
118.
Show HN: Use Wikipedia data for building ML and Information Retrieval tools(github.com)
3 points by crypto5 13y ago|0 comments
119.
by crypto5 13y ago
> Projects where performance is critical are coded in C/C++ , not in java (databases) HBase, Cassandra, VoltDB
120.
by crypto5 13y ago
Bodyshops often cheat DoL classification, by assigning label of System Analyst to i.e. Java Programmer in order to pay less.

More ›