Das Blog @sheerbitchitude - Tumblr Blog

politijohn

happy Valentines day

i… I made some valentines involving politics.

#valentines

yesterdaysprint

The Boston Globe, Massachusetts, March 5, 1924

#WAT

transitmaps

Fantasy Map: Major Rivers of the United States by Theodore Rindos

The early history of the United States is filled with explorers and pioneers making use of rivers as “rapid transit” to quickly travel through otherwise inhospitable or unknown lands. So it makes sense that someone would attempt to show this network in the simplified form of a subway map, and here is one such example by Theo Rindos.

Of course, such a map instantly puts me in mind of Daniel Huffman’s superb series of river/subway maps, but Daniel’s series concentrates more on individual watersheds than the USA as a whole and is a very different beast.

Overall, the map is quite attractive and the comprehensive legend is useful and informative. However, the inclusion of some of the rivers on the map seems somewhat arbitrary to me – most of the “main line” rivers seem to be chosen in order of descending main branch length, but then the Yellowstone River is included instead of a large number of longer rivers. The secondary list of “bus routes” similarly skips quite a few rivers, possibly with the intent of including maps like the Hudson and Sacramento in regions that lack longer rivers.

This does have the result of producing a nicer-looking network which spans the entire country, but at the cost of making sense of the underlying data (which is missing certain longer rivers simply because they couldn’t fit). The inclusion of all three parts of the Platte River is particularly noticeable in this regard as they obviously are there to create a “connecting bus service” between the Colorado and Missouri Rivers.

Also a little odd is Theo’s two-letter abbreviation for each line, which insists on using “R” for “river” as the second letter. This gives us the non-useful designation of “UR” for the Columbia River, as the Colorado has already taken the more obvious “CR”. A better approach would be to use the two-letter state abbreviations for those rivers that share a state’s name, and then more meaningful codes for the remaining rivers. Thus, the Colorado would be “CO”, and the Columbia could then perhaps be “CU” or “CR”.

Finally, Mountain Home, Idaho is some 12 miles distant from the Snake River – the nearest town of any size that actually sits on the river would actually be Glenns Ferry. I can’t comment on the accuracy of other city names further east.

Our rating: Cherrypicks the underlying data a bit to create a comprehensive network, but looks pretty nifty. Three stars.

#maps

aiweirdness

New paint colors invented by neural network

So if you’ve ever picked out paint, you know that every infinitesimally different shade of blue, beige, and gray has its own descriptive, attractive name. Tuscan sunrise, blushing pear, Tradewind, etc… There are in fact people who invent these names for a living. But given that the human eye can see millions of distinct colors, sooner or later we’re going to run out of good names. Can AI help?

For this experiment, I gave the neural network a list of about 7,700 Sherwin-Williams paint colors along with their RGB values. (RGB = red, green, and blue color values) Could the neural network learn to invent new paint colors and give them attractive names?

One way I have of checking on the neural network’s progress during training is to ask it to produce some output using the lowest-creativity setting. Then the neural network plays it safe, and we can get an idea of what it has learned for sure.

By the first checkpoint, the neural network has learned to produce valid RGB values - these are colors, all right, and you could technically paint your walls with them. It’s a little farther behind the curve on the names, although it does seem to be attempting a combination of the colors brown, blue, and gray.

By the second checkpoint, the neural network can properly spell green and gray. It doesn’t seem to actually know what color they are, however.

Let’s check in with what the more-creative setting is producing.

…oh, okay.

Later in the training process, the neural network is about as well-trained as it’s going to be (perhaps with different parameters, it could have done a bit better - a lot of neural network training involves choosing the right training parameters). By this point, it’s able to figure out some of the basic colors, like white, red, and grey: