dumitrue

7 ms·

1.
by dumitrue 10y ago
Apart from the captioning work already mentioned, there's also visual analogy making work by Reed et al. http://www-personal.umich.edu/~reedscot/nips2015.pdf which is pretty exciting (sort of a converse to what yo
2.
Google Brain Reddit AMA Live(reddit.com)
6 points by dumitrue 10y ago|0 comments
3.
Baidu Fires Researcher Tied to Contest Disqualification(bits.blogs.nytimes.com)
14 points by dumitrue 11y ago|2 comments
4.
by dumitrue 11y ago
But the test set has 50,000 images (across 1000 categories). It's not that easy to just "try many times" to get it right.
5.
by dumitrue 11y ago
I don't think OpenCV really solved computer vision to be fair. There's definitely no model out there that can do image-based question & answering as well as a human can, or interpret the contents of an image (parse it, if you
6.
by dumitrue 12y ago
Not really -- he's been an active Google+ poster even before he joined FB ( https://plus.google.com/+YannLeCunPhD/posts )
7.
by dumitrue 12y ago
Just went out: http://arxiv.org/abs/1411.4555
8.
Going Deeper with Convolutions (GoogLeNet Paper)(arxiv.org)
4 points by dumitrue 12y ago|0 comments
9.
by dumitrue 12y ago
Just FYI, the only additional data used by the GoogLeNet entry was from the classification challenge (aka provided by the organizers), hardly something that would make you lose sleep at night.
10.
Computer Eyesight Gets a Lot More Accurate(bits.blogs.nytimes.com)
61 points by dumitrue 12y ago|18 comments
11.
Facebook opens new AI research lab and hires Yann LeCun(plus.google.com)
2 points by dumitrue 13y ago|0 comments
12.
by dumitrue 13y ago
The Los Angeles office is right by Venice Beach, in a beautiful setting. It's a mid-sized office (~500 employees according to http://venice.patch.com/groups/business-news/p/silicon-beach... ) in the same
13.
by dumitrue 13y ago
Yes, though I think on public benchmarks this is still not the case. There's a dog-breed classification problem in this year's Fine-Grained challenge ( https://sites.google.com/site/fgcomp2013/ ) so we&#x2
14.
by dumitrue 13y ago
It's actually already possible to train convolutional network-like models to distinguish between a variety of dogs, cats etc with precision that is pretty much super human. The real problem is getting high-quality training data without
15.
by dumitrue 13y ago
Yes, it'd be surprised if the straightforward implementation from https://code.google.com/p/cuda-convnet/ , run on a GPU with lots of transformations, wasn't the winning entry.
16.
by dumitrue 13y ago
It's possible that the underlying model is just not particularly good at learning from data. 11B parameters is a lot of free parameters to learn -- for instance, the main competitor to that paradigm is the work by Krizhevsky et al., wh
17.
by dumitrue 13y ago
A less ugly alternative, which I much prefer, is to use multiple Chrome profiles -- keeps the worlds separate and the end-result is much cleaner!
18.
by dumitrue 13y ago
Sending a paper for review can help with identifying whether (1) What you're solving has already been solved (2) You have a bug in your solution (3) You can get feedback from the reviewers such that the system is ultimately improved (this h
19.
by dumitrue 13y ago
Sure, but it does depend on the field. In my field (Machine Learning) most (if not all) papers are essentially available online for free (mostly because of very strong actions like http://jmlr.org/statement.html ). There might be some obsc
20.
by dumitrue 13y ago
H1Bs can't vote you out of office so it's easy to be against them, the risk is lower to lose political capital (especially if you frame the discussion in terms of "protecting the jobs for Americans").
21.
by dumitrue 13y ago
I see. Nothing gets to rattle the flame wars around here more than the combo of "is higher education really necessary or can I drop out and build the next tumblr?" and "highly-skilled aliens with PhDs are taking our jobs by accepting lower
22.
by dumitrue 13y ago
There's absolutely no requirement to have a PhD to participate in research (grad students would never get to publish, though I understand this is not what you meant). When you submit an article to a journal or a conference (at least in my f
23.
by dumitrue 13y ago
2 year-old story. Previous discussions: https://news.ycombinator.com/item?id=2019665 [886 days old] https://news.ycombinator.com/item?id=5503996 [46 days old]
24.
by dumitrue 14y ago
Check this out, if you're curious about this line of thought: http://yann.lecun.com/exdb/publis/pdf/farabet-fpl-09.pdf The ImageNet winning model is a essentially glorified ConvNet :)
25.
by dumitrue 14y ago
Basically not true (anymore). If you take the model that is described by Hinton et al and apply it to some vision task for which you generated a lot of data, it will invariably work better than the competition, almost out of the box (yeah y
26.
by dumitrue 14y ago
So actually if you read Geoff Hinton's paper on how they won the ImageNet challenge, they use 2 GPUs. Whether this qualifies as "custom hardware" or not, it's debatable, but it's certainly striking that you can obtain performance that is ab
27.
by dumitrue 14y ago
Why would you ever think that he is being blindly allocated to the position of Director of Engineering? In that position, he'll pretty much be in control of his own destiny at Google.
28.
Ray Kurzweil joins Google(kurzweilai.net)
312 points by dumitrue 14y ago|163 comments
29.
by dumitrue 14y ago
I went to a similar school: one of my fondest memories from college is the raw, unfiltered discussion that took place on those mailing lists. And due to the mail client's ability to thread discussions I could ignore topics that were of no i
30.
by dumitrue 14y ago
Kayak allows you to filter by wifi availability, I've used it many times: after searching go to "more filters" and select "show wi-fi flights only". Note that not even the airline can guarantee that there will be wifi on your flight, as

More ›