Viewing profile — SammoJ
SammoJ
HN member- Joined
- Wed, Jan 05, 2011, 10:13 AM UTC
- HN karma
- 67
- Public activity
- 30 items
- HN profile
- View on Hacker News ↗
About SammoJ
No profile information was provided.
Recent public activity
-
comment
Comment #14600896
Are these numbers for real, and are these all Bay Area? I'm further North and pull no where near these numbers as an ML PhD with a lot of experience now in the tech industry. Have …
-
comment
Comment #13730154
I didn't look too hard but didn't see any similar free services with full DNS management and an API for cache validation. Would love a suggestion here too!
-
comment
Comment #10426137
Seeing a link to the dailymail is a hint that HN may be jumping the shark. This is a UK rag of a paper. Clearly this driver is outside of the recommended highway driving scenario. …
-
comment
Comment #10285067
I was also desperate to see this. And rightly so, here's what looks like a prototype link from above - and it's beautiful! https://www.youtube.com/watch?v=-drmRYsvUaM You can reall…
-
comment
Comment #8856024
Please show a paper where fine-grained vehicle classification in unconstrained images is anywhere near this performance from 20 years ago. You will not be able to, because it wasn'…
-
comment
Comment #8844565
This is brilliant, thank you. On the Musk transcript I found this formatting confusing: "Have you played Kerbal Space Program? What do you think SpaceX uses for testing software?" …
-
comment
Comment #8839484
I don't see anything new here, but for any practioners like myself Scikit-learn and Spyder were the Python tools which finally moved me from a die hard MATLAB junkie. I grabbed the…
-
comment
Comment #8773123
So I had always heard this is the case, and now believe this must be a common misconception. See Table 6 in this file from the Office of National Statistics: http://www.ons.gov.uk/…
-
comment
Comment #8623988
Relevant, very similar paper input/output wise, from our resident karpathy with a detailed discussion in comments: https://news.ycombinator.com/item?id=8621658
-
comment
Comment #8437703
If you read his e-mail he admits he was "completely wrong" http://news.microsoft.com/2014/10/09/satya-nadella-email-to-...
-
comment
Comment #8211252
This would work for 2D games, but 3D projection doesn't allow this. Furthermore in a 3D game there are more than two degrees of freedom.
-
comment
Comment #8181837
Works great in IE11 on an X1 Carbon.
-
comment
Comment #7019751
Were they playing anything or just walking around in the demo? I have one and have tested it out with many people. When people get sick it's usually because they've been in one of …
-
comment
Comment #5478326
Didn't hunt for the source of this one but I did find some true 4K examples here: http://media.xiph.org/video/derf/ CTRL-F 2160p
-
comment
Comment #5436306
I had to drop the www http://visamapper.com/
-
comment
Comment #5379914
Which attributes would you sample from a normal distribution here? I don't see any numerical attributes where this would make sense. One could add weight, height, age etc. and samp…
-
comment
Comment #5359271
Wasn't obvious to me on a first skim through either, but on second viewing all simple variable names are single letters - exactly as one would get through minifying. See for exampl…
-
comment
Comment #5354016
If you grab the resources you can see the lightmaps.
-
comment
Comment #4360781
In a mission to gather data then clearly extra megapixels give you extra data. You cannot argue against that! We aren't talking amateur photography here where the quality of the sh…
-
comment
Comment #4265775
The laser is 6.5KM wide when it reaches the moon. http://en.wikipedia.org/wiki/Lunar_Laser_Ranging_experiment
-
comment
Comment #3711571
Adding complementary features such as edges (I'm guessing this is what you mean by lines) tends to improve the accuracy. It would also be possible to do this in real time. I don't …
-
comment
Comment #3275050
Same for me on FF 3.6.13 Windows7 64
-
comment
Comment #3269454
The deformable part based model is pretty much /the/ state-of-the-art object detection model. There is absolutely no way it should be attributed to "Liu Liu". It's creators are Fel…
-
comment
Comment #2533553
Pedantically a 400% increase would be going from 1 malware app to 5. "The amount of malware is 400% what it used to be" would be 1 to 4.
-
comment
Comment #2469824
I have also used libsvm a lot and can heartily recommend it - but only for non-linear kernels. If you wish to use a linear SVM (which if you aren't familiar with machine learning y…