chudi

10 ms·

61.
by chudi 8y ago
They aren't a hoax, some type of recommendation engines works really well especially the collaborative filtering types. The reallity is that once you have a subset of items that matches your behavior recsys are learning to rank problem
62.
by chudi 8y ago
My experience is that you can't have a generic good recommender system. It all depends on your business but if you just want a mixture between using navigation data and item data lightfm works pretty well for the generic case
63.
by chudi 8y ago
C for the c convention and how to translates to machine code and how you can use this convention to interact with libs. Smalltalk because it's pure, simplistic and there isn't many languages that lives in a image, it's truly
64.
by chudi 8y ago
I use this analogy, you plant seeds in people's minds with your ideas, the ideas grow into their heads and then they realize that it's a good idea. This happens to me a lot with people that doesn't have full expertise in the
65.
by chudi 8y ago
they are hybrid in the sense that gather signals from not just features or user activity, yt paper uses embeddings from search and views, so its more of a mixed model than a pure one content based or a pure collaborative filtering
66.
by chudi 8y ago
I always wondered how do you find the best artwork for the movies, is it multi armed bandits with thompson sampling? In my company navigating politics is always the hard part, the marketing team would love to spam everyone all the time and
67.
by chudi 8y ago
There are basically 3 types of recommender engines: Content Based: If you can represent your products as a vector, you can have a distance between each product, then you have a item-item recommendation. You can use all kinds of embedding to
68.
by chudi 8y ago
We used it to discover the best picture of a product in terms of click trough rate, we also found that the end user really don't want you to change the order of the pictures of their products
69.
by chudi 8y ago
this is a really good answer, at my job we use this approach
70.
by chudi 9y ago
yes, its a weakness that they discuss at the wiki http://wiki.c2.com/?OpenCroquet
71.
by chudi 9y ago
Think object as if it cant draw itself, lets say you have a spreadsheet object that it can draw it self in some window manager so you are updating it with some method calls, etc and you want to share trough the network, what is the differen
72.
by chudi 9y ago
I think that all the people that says that its cool but unpractical are focusing too much on the 3d aspect of croquet. The thing you should really think is that everything is done via objects, so instead of you use json as a sort of rpc, yo
73.
by chudi 9y ago
We use pandas for opening big data files, manipulating the data and then save it to a pickle file or a sql database for use consumption.
74.
by chudi 9y ago
There was also this great ask hn some months ago: https://news.ycombinator.com/item?id=13821217
75.
by chudi 9y ago
We have a number of batch jobs that updates the models every night then a separate api to serve the main app that draws de ui. Sometimes your ml model doesn't need a cronjob and you can update your results near real-time, but the princ
76.
by chudi 10y ago
I'm not from USA. I don't understand why is it dangerous, can't you say that in the USA to your employer? It's just asking for help because you're overwhelmed with your job.
77.
by chudi 10y ago
You get up at the time your contract says and you go to home. Its a matter of expectations, if you usually stay late, people will expect that you will be there. Its like if your boss gives you 10 things to do and you accomplish all of them,
78.
by chudi 10y ago
Check his comments history, the story was already told multiple times
79.
by chudi 10y ago
We use ML for recommendation systems (I work at a Classifieds company)
80.
by chudi 11y ago
It usually takes from 3 to 6 months to know where are the parts of the system that you are using, and also if this is an old codebase maybe longer, your boss knows this, relax, keep learning!
81.
by chudi 11y ago
Graph Theory and Its Applications by Gross and Yellen its a good book.
82.
by chudi 11y ago
maybe you can watch some algorithms classes on youtube? Somthing like this is a good start https://www.youtube.com/watch?v=HtSuA80QTyo&list=PLUl4u3cNGP...
83.
by chudi 11y ago
for databases this is pretty good: http://www.amazon.com/Database-Management-Systems-3rd-Editio... for Oses http://williamstallings.com/OS/OS6e.html cheers
84.
by chudi 11y ago
box 2-3 times a week, running every other day, eat homemade food, sleep all I can
85.
by chudi 11y ago
there is some research in this field from members of my faculty check this links to get started: https://scholar.google.com.ar/scholar?safe=off&es_sm=122&bav... http://ed.fbk.eu/people/sisai&#x2
86.
by chudi 12y ago
The only time that you have to look into the plate of another person its to see if it has enough.
87.
by chudi 12y ago
someone who can code complex algorithms in an environment like a topcoder contest, acm, etc. and can't finish something in the real world.
88.
by chudi 12y ago
Cheap is really cheap, like a poverty vote, but the work, the freedom and the goal is enough compensation to work at a research lab. I don't agree with this compromises that are required to enter the academy but if you want to get into
89.
by chudi 12y ago
Go to the labs that you are interested in, try to fill in the gaps that they cant handle, charge cheap.
90.
by chudi 12y ago
you said that you don't want any flask, but have you looked at Flask-Restless ? Flask just gives you the routing part maybe you can even strip flask out of it and just use wsgi. You can try http://python-eve.org/ too

More ›