jefft255

31 ms·

271.
by jefft255 8y ago
I don't think MOOCs can come even close to be a replacement for a PhD (and from what I understand, neither do you). If a candidate who learned with MOOCs can apply for a job which previously "required" a PhD, then that requir
272.
by jefft255 8y ago
It is not "absolute" as you say, you can get in most research lab with a MS if you are an exceptional candidate. My advice, if you really, really want to be a scientist in a major research lab, is to just bite the bullet and do a
273.
by jefft255 8y ago
I think the claim "scan the whole world down to the detail of a single grain of sand" is hot garbage. SfM or any kind of 3D scanning method doable with cell phone or drone data will not get close to that. You'd need some very
274.
by jefft255 8y ago
Of course they do. Airborne lidar usually works in the near infrared spectrum which is not absorbed by wood. It is however more absorbed by leaves which have a higher water content. My main point is: airborne lidar (which is what OP is refe
275.
by jefft255 8y ago
I think you're mostly right; the only thing I'm more skeptical about is the availability and reliability of high def maps and semantic knowledge as you say. Maintaining such thing will be expensive and I wonder how it will scale c
276.
by jefft255 8y ago
That's not really a good point. When New England/Canada drivers drive through a snowstorm, the vast majority of them get to where they were going just fine. In comparison, basically no self driving car could safely drive 100 ft du
277.
by jefft255 8y ago
You know, there is more to deep learning research than meta-learning/architecture exploration. Sure you can explore the hyper-parameter space faster with 500 GPUs and get yet again a 0.05% better test score on ImageNet (or more I don&#
278.
by jefft255 8y ago
You can buy aftermarket Android Auto headunits: https://www.androidauthority.com/best-android-auto-head-unit...
279.
by jefft255 8y ago
In Canada we have a "good Samaritan law" preventing people from getting sued for trying to help people in distress and messing up.
280.
by jefft255 8y ago
It highly depends on what exactly you want to integrate on which robot. Stuff like control algorithms can be dependent on the type of robot used. Also, a lot of this stuff is already present in ROS as a downloadable package.
281.
iRobot CEO interview on the future of smart homes(theverge.com)
3 points by jefft255 9y ago|0 comments
282.
by jefft255 9y ago
I'm not OP, but my take on this is that it would be quite difficult to design a DNN working directly with vector graphics. I'd say you should rasterize your image before inputting it to the network. Same with outputs. CNNs would d
283.
by jefft255 9y ago
There are many projects using deep learning with lidar. Google PointNet, PointNet++, also https://www.youtube.com/watch?v=UXHX9kFGXfg . These are all much newer than 2D CNNs and I don't know if it works well enough to
284.
by jefft255 9y ago
Just to be clear, neural networks typically aren't used to work with lidar in self driving cars. NN for point clouds are still at the research stage.
285.
by jefft255 9y ago
I thought GMOs were illegal in the EU, of which France is a part of
286.
by jefft255 9y ago
I know right? Back in the day my iPhone 4s would almost always shut down on my way to university in winter. It was a 30 minute walk in cold, windy Quebec weather, but my S7 does not have this problem. Why is this news?
287.
by jefft255 9y ago
The "AI" algorithm used in the paper is far from a black box. It's logistic regression, which is extremely well understood and has been used by statisticians and scientists for decades.
288.
by jefft255 9y ago
It's actually a stronger assumption than simply being continuous
289.
by jefft255 9y ago
Most universities in Canada have the same thing. In my university you can do a masters "avec mémoire" (with thesis) which is four semesters of full time research.
290.
by jefft255 9y ago
C&C Tiberian Sun had that in the 90s. You could tweak a lot of parameters as well!
291.
by jefft255 9y ago
You rarely need to pay if you work at a university. When you're on the school network or VPN there's a way to bypass the paywall because the uni usually buys subscriptions for most journals. That does not work if you are not worki
292.
by jefft255 9y ago
I am a math double major doing a research masters in ML/CV right now, and I know plenty of pure CS majors who are doing just fine. The math that 95% of ML scientists use is not that hard to grasp. Sure, when they encounter functional a
293.
by jefft255 9y ago
It's more often a question of whether it's worth using a robot instead of humans. I'm pretty sure it's already possible to automate much more than what's already automated in a car production line. Sometimes, for a
294.
by jefft255 9y ago
In my lab there is a masters student working on monitoring deforestation for palm fields in Indonesia using Google Earth Engine, which is similar to Earth on AWS. There is a whole scientific field devoted to analysing this kind of data: rem
295.
by jefft255 9y ago
I totally agree with your last point, and I wish more university acknowledged this when deciding who teaches classes. My favourite CS course was taught by a first year MS student. For most universities though, letting an undergraduate teach
296.
by jefft255 9y ago
Also, there's something I don't get: how does an undergrad ends up teaching a full blown course at Stanford? Especially considering she doesn't seem to be some kind of super genius, only a probably excellent student. I'm
297.
by jefft255 9y ago
If they take an Operating Systems course (which, I hope, every CS grads does) then they certainly get taught C. That was the case for me and I only graduated last year. Plus, most CS grads learn C++ which means they can easily get started w
298.
by jefft255 10y ago
This kind of attitude is what's wrong with the software development world.
299.
by jefft255 10y ago
Yes it really is underappreciated. As quoted by other comments, "Most businesses think they need advanced ML and really what they need is linear regression and cleaned up data". A significant portion of businesses currently invest
300.
by jefft255 11y ago
It depends what kind of job you've got now. If it's technical, let's say engineering or science, data analysis etc... , then you can learn how to code on your own and build something that you can show to potential employers.

More ›