Category: Uncategorized

The signal I’m looking for makes up about 0.1% of the sequence. This isn’t a huge problem, as filtering out the parts of the signal that don’t change and passing the parts that do is what a filter does. Setting up the filters is tricky but solvable. Tricky but solvable means time-consuming, and that’s the issue.

Happy Labor Day, everyone. Celebrate by working!

If you think of programming as writing in the English passive voice, you’re most of the way there.

Most of the rest is A) complaining about the passive voice B) finding other people who also complain about the passive voice C) really just whining D) getting over yourself or E) repeating A-C for a while. The rest rest is semicolons.

A = myfunc(b);

A is myfunced b.

If you really want a big brain moment, a script is one horribly run-on sentence with way too many clauses. Think patent-claim level run-ons.

Most exceptions are fairly straight forward. A lot of operators are imperatives.

A++;

Increment the A.

Pronouns basically don’t exist, but it’s very easy to make up proper names.

Honestly, I really don’t understand why basic computer programming isn’t in the foreign language department. It’s just super picky grammar with a reader that won’t understand/give-any-benefit-of-the-doubt ever.

The reason Starbucks exists is I can sit down, drink a coffee, and do nothing for thirty minutes.

In the US most restaurants have one of two purposes: serve food or serve alcohol. Those are their purposes. You eat, then you get the hell out. Or you have a few beers and sit around. There are fairly few establishments where in the early afternoon I can drink coffee, not beer, and do a little nothing. Sometimes I want to read a book. Sometimes I want to play on my phone outside the apartment. Sometimes I want to sit at a table and space out. I will achieve the Zen of mental absence. At 2 pm I don’t necessarily want a beer, but restaurants are focused on serving food and turning tables.

Coffee shops fill a vacancy, with Starbucks being the leader. People whine about burned beans and overpriced coffee, but those whiners are missing the point. I’m buying an hour of table space in a reasonably quiet facility outside my apartment. The cost of providing that table space is charged in coffee, and comes to three to six bucks a cup. AC, a swept floor, a reasonably comfortable chair: those add up to five-dollar coffee at Starbucks quality. Whatever. That’s the price to sell.

Read something on McArthur Binion, artist in Detroit, and what left me the most curious was an off-hand comment that he used to drive taxis to make ends meet.

I am no art buff and make no pretensions of expert criticism. He clearly has a lot going on upstairs. I wonder what taxi driving was like. Was his head full of ideas? Did he compartmentalize? Was every fare two more hours in the studio to accomplish something? Did he zone out and veg, or was it a constant battle of hopes, curiosity, wonder, and irritation?

Am ded

The doctor told me stop lifting weights until I blackout.

I’ll show her!

*dies*

I wish I had more time in this city.

I do; I just wish I had more.

Heading North into Georgetown. The valley north of this vantage drops significantly, and if the ground wasn’t in the way, Georgetown would be dead ahead and slightly to the right.

So I got the numbers.

The morning after in terms of leaving my computer up and running all night, btw.

The numbers exist. They’re even worse than I expected, and I half-heartedly expected fire. But that’s actually not so bad, because I can start poking at them now.

I’m using somewhere in the vicinity of 12,000 samples, each one 20,000 data points raw. I downsample by factors of 2, 3, 4 or whatever so they don’t throw memory errors, and up-and-downsample batch sizes for accuracy. I was really surprised that didn’t make a whole lot of difference.

My validation dataset may be the issue, so that’s the next setting to change. After that, I’ve got a dozen ways to manipulate GradientThreshold, batch size, weights, and momentums, etc. I can downsample by factors of 10, 100, or more and throw huge minibatches around. If nothing changes the results, I have bad data. The data doesn’t clearly indicate the stuff I’m looking to measure.

However, I know the signal is in there. That makes this somewhat easier, because I know the problem is solvable. It just takes longer than one might think.