ramanan

True. I’m no expert in this area, but there are a lot more factors too - including territorial range and even different camera locations.

This catch and release pattern is quite common in many wildlife surveys.

For example, when estimating tiger populations, instead of painting a white dot, rangers set up camera traps to take photos. Then they can use stripe patterns as a signature for subsequent re-appearances. Quite an interesting intersection for image AI with statistical counting methods.

This downward trend has long been in the works.

The current government came into power riding on a wave of popularity, largely driven by their "IT cells". Now comfortably in power, they are probably dismantling the path, to avoid others who might try to follow.

The government is not just in a stand-off with big-tech companies. It continues to view its own citizens as adversaries with frequent internet shutdowns [1].

[1] https://internetshutdowns.in/

Spotting a distressed swimmer is possibly much harder than most people expect.

There is a great website where you can try your spotting skills yourself [1].

[1] http://spotthedrowningchild.com/

minutephysics had a nice video about the mathematics that go behind piano tuning too.

"Why It's Impossible to Tune a Piano" https://www.youtube.com/watch?v=1Hqm0dYKUx4

Got it. I was looking for input on how generalizable (the ability of weights to change/adapt) when the training labeled data is 100x smaller than the initial pre-training dataset?

Also, I don't understand the need to be so defensive though and the relevance between my employer and my post?

I understand that you do mention the pre-training / transfer learning approach clearly, but isn't it disingenuous to claim that you provide better performance based on (only) 100 labeled examples, when the pre-training dataset (Wikitext-103) actually contains 103M words?

Well, one easy way when using Ubuntu-like distributions is to simply place your `/home` folder on a separate (persistent) EBS volume [1].

With a few on-boot scripts to attach-volumes / start-containers, it should be fairly easy to get going as well.

[1] https://engineering.semantics3.com/the-instance-is-dead-long...

there are a number of similar paradoxes that arise when considering infinities of different sizes!

Infinity Paradoxes - Numberphile - https://www.youtube.com/watch?v=dDl7g_2x74Q

Until this interesting video [1] by Tom Scott, I had never considered sand as a non-renewable resource. But now, it is clear that we now have one more resource that we might run out of.

[1] The World Is Slowly Running Out Of Sand - https://www.youtube.com/watch?v=kMLYLcniXIc

Aah yes, every change breaks someone's workflow: https://xkcd.com/1172/

Reminds me of one of my all-time favorites -

The 500 mile email

https://www.ibiblio.org/harris/500milemail.html

Bicycles are surprisingly stable as long as they are able to maintain some momentum.

minutephysics had a great video about this too:

How Do Bikes Stay Up?: https://www.youtube.com/watch?v=oZAc5t2lkvo

It wouldn't be that easy. The drone would still need EM windows for its other requirements, possibly communication and navigation.

OP here, the post was really meant to invite discussion.

I would love to hear more from people deploying machine learning techniques at their projects, teams and companies.

For those interested in the use of microwave links for HFT, there was an interesting post [1] a while back and some discussion on HN [2] too.

[1] https://sniperinmahwah.wordpress.com/2014/09/22/hft-in-my-ba...

[2] https://news.ycombinator.com/item?id=8354278

You might be interested in bro pages then!

http://bropages.org/

Reminds me of a similar article covering the work of Douglas Hofstadter, the author of GEB :

The Man Who Would Teach Machines to Think

http://www.theatlantic.com/magazine/archive/2013/11/the-man-...

Proactively thinning selective regions of forests and controlled burns seem to be the route forward to prevent (rather than react to) large-scale fires.

This short video from MinuteEarth, while not going into much detail, does have a lot of good references

https://www.youtube.com/watch?v=NX1xnWPSjKg

A related video on the numbers behind piano music, on why it is mathematically impossible to perfectly tune a piano :

https://www.youtube.com/watch?v=1Hqm0dYKUx4

I remember seeing this on Numberphile a few weeks ago [1].

For anyone interested in this, I suggest taking a look at hyperbolic baseball & golf [2] too !

[1] https://www.youtube.com/watch?v=4KHCuXN2F3I

[2] https://www.youtube.com/watch?v=u6Got0X41pY

I think that you are referring to LTCM ( https://en.wikipedia.org/wiki/Long-Term_Capital_Management ) which blew up in the late 90s. For others interested, I would strongly recommend the book "When Genius Failed" ( http://www.amazon.com/dp/0375758259 )

Related link : The Mercator Puzzle : https://gmaps-samples.googlecode.com/svn/trunk/poly/puzzledr...

A nice way to get a feel for the distortions.

CGP Grey! I came here to post the same link. His entire series is very informative. https://www.youtube.com/playlist?list=PL7679C7ACE93A5638

Perhaps what throwaway8198 is referring to are the practical consequences of Maxwell's laws as formulated by an American team in the 1960s [1]. For a thorough, more recent study, I highly recommend [2].

[1] https://en.wikipedia.org/wiki/Chu%E2%80%93Harrington_limit

[2] http://web.eng.ucsd.edu/ece/courses/ECE222D/Spring2013/Lectu...

After having read the PRL paper [1], I am not really sure why they left out gain (and pattern) measurements for the antennas. In my experience, radiation efficiency is only one part in determining the received signal strength. The impedance match is also crucial in determining the effective gain.

The figures in their paper also seem to demonstrate strong resonant coupling effects, which do not always translate to effective radiation.

[1] http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.114...

and what number would you choose ? waitbutwhy had an interesting discussion on this a few weeks back: http://waitbutwhy.com/table/how-long-would-you-live-if-you-c...

My favorite story about the application of statistics to the second world war, is the `German Tank Problem`. Statisticians' estimate of German Panther tanks proved to be much more closer to the real number than conventional intelligence estimates. https://en.wikipedia.org/wiki/German_tank_problem

The goals of the new research center seem close to what Douglas Hofstadter (author of GEB) has been working on for all these years.

A related HN post from a few weeks ago: https://news.ycombinator.com/item?id=6605015

The situation is still evolving. The article has been updated since the initial HN posting and the title now reads 'Snowden asylum still under review, stays in airport for now - lawyer'.