squigs25

8 ms·

31.
by squigs25 11y ago
Good fresh home made turkish delight is actually awesome. If you ever come across it, try it. The pre-packaged stuff is mostly too sweet and not that good.
32.
Opportunity: Doctors Complain About Electronic Health Record Technology(youtube.com)
1 points by squigs25 11y ago|0 comments
33.
by squigs25 11y ago
Enjoy Garry! I was just in the south of France, it was a very different world and one of the best trips I've taken. Let me know if you're looking for any recommendations.
34.
Australia Deploys Sheepdogs to Help a Penguin Colony Back From the Brink(nytimes.com)
2 points by squigs25 11y ago|1 comments
35.
by squigs25 11y ago
Thought this was going to be about Adaboost on Apache Spark, but it is not!
36.
Hacker News Top Words(github.com)
1 points by squigs25 11y ago|0 comments
37.
The Angola Prison Rodeo(en.wikipedia.org)
2 points by squigs25 11y ago|0 comments
38.
by squigs25 11y ago
Dimensionality reduciton is correct - the cosine similarity on a character/word level really just provides a metric for measuring the "anagramness" of two words. But "Mary" and "Army" would have a perfect
39.
The Craigslist Lawsuit(3taps.com)
209 points by squigs25 11y ago|122 comments
40.
by squigs25 11y ago
For these reasons I have been using ear muffs (similar to the ones you would use for lawn mowing). Currently I am using Howard Leight Sync. It's basically an "over the ear" ear plug when it's not on. You can also use the
41.
by squigs25 11y ago
On a tangent, why is this problem soooo pervasive? I would say the large majority of people I work with fall into the over-confidence bucket. Perhaps it's that people in tech are smart, and usually right when they find themselves in an
42.
by squigs25 11y ago
Agreed. I would say this answer closely resembles the Socratic method, which is the frame of mind I try to maintain when dealing with the over confident souls out there.
43.
by squigs25 11y ago
Was there a training set made available and distributed? I got the impression that there was not.
44.
by squigs25 11y ago
In a contest like this, I would expect (and I could be wrong) that there is a final validation set not accessible until "game day". Thus, overfitting to a training data set of any kind would be detrimental and degrade model perfor
45.
by squigs25 11y ago
If they are building an algorithm which overfits to training data, their algorithm will lose performance on in the actual validation test. It is NOT like looking at leaked answers. It's like re-reading the textbook more than everyone e
46.
by squigs25 11y ago
This is clearly cheating. HOWEVER, the goal of these contests should be to promote the most accurate and powerful image recognition algorithms that will transform the world as we know it. Limiting access to training data makes it more diffi
47.
by squigs25 11y ago
() is not a tuple. (a,b) is a tuple or (a,) would also be a tuple.
48.
Moore's Law Turns 50(nytimes.com)
1 points by squigs25 11y ago|0 comments
49.
by squigs25 11y ago
then again, your username is Rainymood :-)
50.
Google Code Challenges(google.com)
1 points by squigs25 11y ago|0 comments
51.
by squigs25 11y ago
I believe this algorithm must (partially) base the sentiment decision on the first name of the target. A couple of my friends with an unusual first name did not get a result, even though I know they are on LinkedIn. Also, very common names
52.
Being a night owl might be bad for your health(well.blogs.nytimes.com)
1 points by squigs25 12y ago|0 comments
53.
Lean Start-Up Methodology Simplified(thinkapps.com)
2 points by squigs25 12y ago|0 comments
54.
Big Oil Pressured Scientists Over Fracking Wastewater's Link to Quakes(bloomberg.com)
4 points by squigs25 12y ago|0 comments
55.
Science Hostel(en.wikipedia.org)
3 points by squigs25 12y ago|0 comments
56.
China's Three Gorges Dam has slowed the Earth's rotation by 0.06 microseconds(businessinsider.com)
2 points by squigs25 12y ago|0 comments
57.
by squigs25 12y ago
scikit learn is a great package, but it's bizarre that people haven't heard of it before?
58.
by squigs25 12y ago
I'm really pumped about the update to GIN indexes, and the ability to concurrently update materialized views. Both enhancements are huge for the postgres ecosystem, and especially for productionizing postgres databases.
59.
by squigs25 12y ago
If anyone ever needs authentication in their flask/django app, I highly recommend authomatic
60.
Authomatic: user authentication in Python made simple(peterhudec.github.io)
3 points by squigs25 12y ago|1 comments

More ›