vinn124

5 ms·

1.
by vinn124 8y ago
for a good introduction to order and complexity (including intelligence) arising from nothing, read "the origins of order" by stuart kauffman.
2.
by vinn124 8y ago
who cares?
3.
by vinn124 8y ago
> At the core RL is just updating a table of values, and then using function approximation (aka, machine learning) for more complex cases. this seems to be a common assertion about ml. other refrains include "ml is just matrix multi
4.
by vinn124 8y ago
particularly when most microsoft investors value microsoft stock for its earnings potential, as opposed to growth/revenue potential.
5.
by vinn124 8y ago
> Increased accuracy per unit wallclock time is what you want. yah, especially for a framework for distributed learning!
6.
by vinn124 8y ago
> 3) Test that the solution works on toy examples, like MNIST, simple block worlds, simulated data, etc. youre right: mnist, imagenet, etc are toy examples that do not extend into the real world. but the point of reproducible research
7.
by vinn124 8y ago
> What are the use cases for adding yet another layer to the stack? in my limited experience with horovod, horovod is most useful when youre running large clusters of workers/ps. in those situations, you typically have to manually f
8.
by vinn124 8y ago
> Usually you throw everything and see what sticks. most practitioners start with the simplest possible learner, then gradually, and thoughtfully, increase model complexity while paying attention to bias/variance. this is far from a
9.
by vinn124 8y ago
> Many losses which don't seem differentiable can be reformulated as such... agreed, especially with policy gradients. > If the dimensionality is small, second-order methods (or approximations thereof) can do dramatically better
10.
Ai and compute (since 2012)(blog.openai.com)
1 points by vinn124 8y ago|0 comments
11.
by vinn124 8y ago
not exactly what you asked for, but if youre looking for a gentle academic introduction to the intersection of ai and games, i recommend [ AI Researchers, Video Games Are Your Friends!]( https://arxiv.org/abs/1612.01608
12.
by vinn124 8y ago
the quality and preciseness of this answer is why i read hn every day.
13.
by vinn124 8y ago
> Why should we expect there to be any mathematical foundation to this stuff? i would be surprised that "this stuff" would be exception to the unreasonable effectiveness of mathematics. mathematics underpins virtually every obs
14.
by vinn124 9y ago
youve missed the point. the point is: anything complex can be dismissed as "just x,y,z" if you dont appreciate the massive body of work behind it. i made that point because OP observed that "ml is just affine transformations&
15.
by vinn124 9y ago
> Then my next question will be how can I trust something that I don't understand. ive always found this mentality silly. do you understand aerodynamic principles? laws of electromagnetism? mendelian genetics? or information theory?
16.
by vinn124 9y ago
how does this solve anything? if a simple decision tree could predict the outputs of more complex deep nets, why not use the decision tree in the first place? also, what do you do when a decision tree isnt powerful enough, as in the case of
17.
by vinn124 9y ago
it was the ability of steve jobs to create an organizing principle (and its corresponding organizational structure) that deserves credit for apple's success.
18.
by vinn124 9y ago
> It's just alternating layers of affine transformations and nonlinearity with lots of tricks and improved routing. and a computation is just 0s and 1s, with lots of if/then statements.
19.
by vinn124 9y ago
pytorch, not tf, seems to be winning the hearts and minds of ml researchers. then again, things change quickly.
20.
Programming for Computations: An Introduction to Numerical Simulations(hplgit.github.io)
2 points by vinn124 9y ago|0 comments
21.
by vinn124 9y ago
> If you're not waymo you probably use ROS. is this true? ros works well for prototyping, but is it actually used in prod?
22.
by vinn124 9y ago
> I am curious to know what major innovations in search engines happened since the page rank algorithm, or were there only incremental improvements? a ton has happened! since pagerank, theres been a ton of advances around nlp that has ch
23.
by vinn124 9y ago
> Note that the article has a comment by Yann LeCun (hopefully it's not an impersonator). i wouldnt be surprised if it was lecun. colah's illustrations on nonlinear transformations have made it into several lecun papers, includ
24.
by vinn124 9y ago
agreed
25.
by vinn124 9y ago
that is true: popular rl techniques (eg policy gradients) are very similar to "vanilla" supervised learning techniques and architectures, but they are unsupervised in the sense that they required zero human input. alphago zero i
26.
by vinn124 9y ago
sounds like a job for [adversarial networks]( https://arxiv.org/pdf/1406.2661.pdf )!
27.
by vinn124 9y ago
> The test of believability yep! that is the fundamental limitation of adversarial networks. theres no good measure or "loss", as it's highly subjective.
28.
by vinn124 9y ago
> but isn't that the whole point of GANs? not quite, but youre on the right path. think about it this way: you (the generative model) are trying to predict a unit gaussian, which is just a fancy way to say bell curve. you get +1 if
29.
by vinn124 9y ago
deployment is not the only reason. distributed ml (not deepmind scale but we're still talking about several instances, usually for something like hyperparameter tuning) is fairly common in practice. without docker, id waste most of my
30.
by vinn124 9y ago
what a lovely project - thank you for sharing

More ›