hey /sci/ what are you currently researching?
I'll start:
>Applications of simulated-annealing global optimization to develop rock mechanics models
>pic related
>Induction heaters
>Higgs Boson
>>7678558
At the moment I'm in more of a research stage, and it seems that this is a realitivy unexplored technology. An Induction heater can be used to join/weld metal together, or can be used to completely liquidfy metals.
>>7678566
Kek
>>7678512
Gauge fields in electromagnetism
>>7678512
How human visual perception changes depending on the ambient light levels.
>>7678558
>Anything specific?
Yeah, looking at the coupling to Tau leptons with the new data.
Machine learning. To be more specific, recurrent neural network.
It's the only network right now with a short term memory. However, it is really hard to train this little shit.
Normal sparse/dense deep NN are able to work well with most basic tasks though.
>>7678590
Interesting! Any general rules of thumb you could share just as a teaser?
>>7678606
Nice! I really want to get into machine learning more seriously. I use neural networks every now and then, and kind of pretend to understand how they work, but I don't think I can code one from scratch. I'm a geophysicist with a background in inverse theory, but I'm used to more physics-based parameter estimation. However, machine learning seems to be some damn powerful that it should at least give us some solutions that can push the development of physics-based theories in the right direction. But, having said all that, I don't know much about machine learning... Any suggestions to getting started? I've heard Andrew Ng's coursera Machine Learning course is excellent.
>>7678628
>Interesting! Any general rules of thumb you could share just as a teaser?
Not too much because we applied for a patent recently. In a very condensed way, our perception is tied to what the lowest light intensity increment is that humans can perceive. If you know what the increment is with no ambient light (that was measured and is expressed as a set of contrast sensitivity functions or CSFs), how it changes with increasing ambient light and how to extrapolate that data from simple patterns to complex images, you can create a model that will predict it for an average observer.
>>7678628
Andrew Ng's ML course is a great start. The pace is slow but all the stuffs he teaches is all necessary.
If you want to learn more about NN, get Yoshua Bengio's book.
>>7678512
quorum quenching molecules
>tfw microbio
The synthesis of LiFePO4 in ionic liquids.
>>7678645
>The synthesis of LiFe
Abiogenesis confirmed.
>>7678645
>Purple back lighting in a forest
Only in anime.
I give mice heart attacks and then try to help them recover using stem cell therapies.
So far... none have survived :(
Coanda lifters
/sci/ thinks it's all bunk however.
![B-YPraNCQAAjHOL[1].jpg](https://i.imgur.com/eWZrO1l.jpg "B-YPraNCQAAjHOL[1].jpg")
>>7678747
;_;7
>>7678762
Here we go with the disbelievers.
>>7678823
But Coanda isn't even that hard to lift. She weighs like 135 lbs.
>>7678751
bonsai tornados?
>>7678747
Shit son...
