Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-06-18

| Channels | #pdl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
14:11 vicash left #pdl
17:38 vicash joined #pdl
17:40 spintronic joined #pdl
17:42 spintronic So I'm wondering. Are all of you astronomers?
17:45 sivoais joined #pdl
17:46 spintronic So I'm wondering. Are all of you astronomers?
17:52 jberger spintronic: I know some people are still
17:53 jberger but there is usage all over the spectrum
18:19 sivoais joined #pdl
18:25 vicash spintronic: what are you interested in using PDL for
20:04 spintronic machine learning and data analysis
20:05 spintronic so far I've found these examples: https://github.com/navoj/PDL-ML
20:06 spintronic Has anyone created a neural network library with PDL?
20:11 mohawk not that i'm aware of
20:11 mohawk but have you googled?
20:11 mohawk http://search.cpan.org/~zpmorgan/AI-Nerl-0.03/lib/AI/Nerl/Network.pm
20:12 mohawk yes they have
20:13 mohawk https://metacpan.org/release/AI-Nerl # better url
20:13 jberger zpmorgan used to be around here on IRC
20:13 mohawk cor
20:14 mohawk pdl seems highly suitable for NN with its badly-named "threading" and propagation features
20:26 vicash there is also a PDL::Stats somewhere
21:06 mohawk yes there is
21:06 mohawk MAGGIEXYZ gave a very interesting talk on it
21:06 mohawk it's on youtube
21:39 pdurbin nice. I just realized that my http://blogs.perl.org/users/philip_durbin/2011/05/john-siracusa-discusses-perl-on-hypercritical.html was linked to from https://changelog.com/133/ :)
21:54 spintronic Thanks I've found the AI-Nerl module as well while searching. It's a good start for me to expand on.
21:56 jberger does anyone else read AI-Nerl as Al Neri ?
21:57 spintronic Is Mark Baker ever in this channel?
21:57 jberger Mark Baker ... no
21:58 jberger why do you ask?
21:58 spintronic I'm trying to find more information about his simulations using PDL.
22:00 jberger I know he had a website and a bunch of youtube videos
22:00 jberger but TBH, his code ... lack a sense of ... well ...
22:00 spintronic yeah plus he doesn't post much of his code.
22:01 jberger I'm trying very hard to be diplomatic, but perhaps you might do better to find other examples
22:02 spintronic still looking
22:05 spintronic jberger: looks like you have a few physics simulations as well ;-)
22:06 jberger I do
22:06 jberger I don't use PDL for many of them, just because mine didn't fall in its niche
22:07 jberger I did use it for image processing a lot, but nothing worth sharing
22:07 jberger it was very purpose built
22:07 spintronic oh ok
22:07 jberger I did use PDL as an interface to plotting though
22:07 jberger PDL::Graphics::Prima and PDL::Graphics::Gnuplot primarily
22:08 jberger https://metacpan.org/source/JBERGER/Physics-UEMColumn-0.901/examples/Cooke.pl#L70
22:08 spintronic yup a big improvement over the old PGPLOT interfaces.
22:11 spintronic You wouldn't have to have any models for heat propagation through semiconductor surfaces by laser pulses would you?
22:11 spintronic *happen
22:12 jberger hahahah, you aren't serious right?!
22:13 spintronic of course I am
22:13 jberger spintronic: https://www.osapublishing.org/josab/abstract.cfm?uri=josab-29-4-543
22:13 spintronic right now I'm blasting metal traces with a laser, but we are trying to see if it's possible to just melt the metal traces for other applications.
22:15 spintronic haha beautiful thanks jberger :-)
22:15 jberger just fyi, the math is scary, like really scary
22:15 jberger though it might be simplified if the target isn't a sesam
22:16 spintronic they wouldn't pay me if it were easy ;-)
22:16 jberger because that has nonlinear response AND it affects the laser performance (since it is an end mirror in the cavity)
22:17 spintronic oh this is modeling the heating of the laser cavity itself huh?
22:18 jberger in ultrafast lasers you employ a semiconductor mirror on one of the ends
22:18 jberger and as it heats it changes the laser properties, which changes the heating etc etc
22:18 jberger this was about modeling that
22:18 jberger without the feedback the math probably simplifies
22:21 spintronic ah ok.
22:22 spintronic I've been running experiments where we have a metal-insulator-metal stack and we are trying to use a green laser to weld the stack together.
22:23 spintronic I also have a wavelength in the infrared available.
22:26 spintronic pulse rate is 1 HZ, so I don't think we have the right tool for the job.
22:27 jberger spintronic: I don't know anything about laser welding, so I can't help you there
22:28 spintronic I need to read up on basic optics. I don't even know what the energy output of this system is.
22:32 spintronic I think at a rate of 1 HZ the surface probably cools off instead of heating up to melt the materials. I just need the thermal conductivity of the surface, the amount of energy per shot, and the heat capacity for phase change.
22:33 spintronic a simple 1-D heat transfer model would probably be sufficient to show this setup won't work.
22:34 spintronic gtg, my LabView app just locked up. Need to reboot.

| Channels | #pdl index | Today | | Search | Google Search | Plain-Text | summary