Monica

avian socializing in the 21st century

2023-04-26T21:47:06Z

How nifty!

Parrots are social creatures. However, most pet parrots are singletons. They get lonely and sometimes that leads to destructive behavior.

From the Smithsonian:

Once the birds had learned how to initiate video interactions, the second phase of the experiment could begin. In this “open call” period, the 15 participating birds could make calls freely; they also got to choose which bird to dial up. Over the next two months, pet parrots made 147 deliberate video calls to other birds. [...]

For starters, they found that the parrots took advantage of the opportunity to call one another, and they typically stayed on the call for the maximum time allowed during the experiment. They also seemed to understand that another live bird was on the other side of the screen, not a recorded bird, researchers say. Some of the parrots learned new skills from their virtual companions, including flying, foraging and how to make new sounds. [...]

The birds forged strong friendships, which researchers measured by how frequently they chose to call the same individual. Parrots who initiated the highest number of video calls also received the most calls, which suggests a “reciprocal dynamic similar to human socialization,” per the statement.

The article links to this ACM paper. Yes, ACM-CHI, meaning it's from a technical conference not an animal-behavior conference. (Also, I guess this stretches the boundaries of the 'H' in CHI, which stands for Computer-Human Interaction, or at least did the last time I attended that conference.)

comments

TIL: equinox, kind of

2018-09-23T19:30:28Z

Yesterday was the equinox, but I couldn't help noticing that sunrise was at 7:07AM and sunset was at 7:16PM. That stretches the definition of "equi" a bit. Looking ahead, the day won't be within a minute of 12 hours until September 25 or 26. (One's a minute longer, one's a minute shorter.) So off to Google I went.

There are two things going on, it turns out. The first is that the equinox is relative to the center of the sun, but we count sunrise and sunset from when the top is visible. But that only accounts for 2.5-3 minutes at my latitude.

The bigger factor is atmospheric refraction: after the sun has actually set (all parts past the horizon), or the reverse in the morning, you can still see the sun. What? Yeah, apparently you can look westward at sunset and see "the sun" even though the sun is not in your line of sight; light bends. This effect varies with atmospheric conditions, but is usually good for about six extra minutes of day.

I said that I won't see a 12-hour day here for a few more days. Apparently that effect gets stronger as you move toward the equator; this site says at 5 degrees North that date isn't until October 17. It also says the day is never exactly 12 hours at the equator, when I thought the equator was the one place where you had reliable 12-hour days all year. Today I learned.

I wonder -- because I'm the sort of person who wonders about stuff like this -- what the effect is in halacha, Jewish law. The day starts at sunset, but when beginning Shabbat we add some extra time just to be safe -- 18 minutes in most communities. That's l'hatchila, what you should do from the outset, but b'dieved, after the fact, if you cut into the 18 minutes with your preparations, it's ok because it's not actually sunset yet. Except... maybe it is? If you have a bad week and light candles two minutes before (nominal) sunset -- when you can still see the sun in the sky, except it's not there -- have you kindled fire on Shabbat? Or do you go by what you can see anyway? I plan to ask this on Mi Yodeya if it's not already there, but first I have to finish Sukkot preparations.

Update: asked.

comments

view from a distance

2017-08-21T23:13:03Z

We didn't travel to see the total eclipse. Here in Pittsburgh we got just over 80% coverage, so some coworkers and I went to the roof of the parking garage armed with minimal tools to see what we could see. The pinhole cameras were mildly interesting but low-res; none of us had thought to bring interesting things like collanders to make eclipse art on the pavement. One of the other people there lent us glasses for a few minutes, which was nice.

The view through the glasses was very neat -- just a sliver of sun. The picture I took through the glasses shows a much fatter sliver than was really there. I think the yellower sliver inside the larger orange-yellow sliver might be real and the rest bleed-through or something:

It was still pretty bright and sunny out, even with only 20% of the sun directly visible. It's not like you get dusk-quality lighting. I did notice that it wasn't *as* bright as usual; in particular, not only did I not have to squint like I normally do when outdoors on bright days, but I could even *take off my glasses with transition lenses* and not have to squint. That was pretty neat!

I took a quick photo without the glasses (I figured my phone could handle a very brief exposure). It shows no occlusion. Is that what people foolish enough to look with the naked eye would see? So they'd endure vision loss for... not a whole lot of anything?

These people improvised a pinhole camera that worked better than the one I'd prepared in advance:

comments

astronomical puzzle

2017-06-30T23:39:14Z

Unless you're on the equator, neither the earliest sunset nor the latest sunrise of the year is on the winter equinox (source). In Pittsburgh, the earliest sunset is usually around December 9 or 10. People who keep Shabbat tend to notice this.

This happens because apparent solar time doesn't line up exactly with mean solar time. The day isn't consistently (or exactly) 24 hours long, and "noon" usually isn't exactly 12:00. Plus there's some shift because of latitude. Fine.

I wondered why there wasn't a corresponding effect at the summer solstice, and played around with this slider to map it out. There actually is an effect, but it's much smaller -- the earliest sunrise was parked at the same time (rounded to the minute) from June 10-19, and the latest sunset is parked at the same time from June 23 - July 1. So in the week or so surrounding the solstice there's barely any change, while in December the boundaries more more visibly. The latest sunset is June 23 which is barely past the solstice, but it's also July 1 (and every day in between of course). And the earliest sunrise is only a couple days before, but also a week before. So what I notice is "earliest sunrise June 19, latest sunset June 23", even though those bands are wider. In the winter, on the other hand, sunset has been creeping later for a week and a half when you get to the solstice.

I guess this, too, is because of latitude, but it's still not intuitive to me. I wonder what's still wrong with my mental modeling.

comments

purim science?

2017-03-13T14:21:53Z

One machine-learning technique is to pit evolving neural networks against each other in cage matches and then learn from the results. This is called Generative Adversarial Networks (GANs).

At yesterday's Purim festivities somebody described the following cutting-edge research, and I remembered just enough keywords to be able to find the paper later:

Stopping GAN Violence: Generative Unadversarial Networks
Samuel Albanie, Sébastien Ehrhardt, João F. Henriques

While the costs of human violence have attracted a great deal of attention from the research community, the effects of the network-on-network (NoN) violence popularised by Generative Adversarial Networks have yet to be addressed. In this work, we quantify the financial, social, spiritual, cultural, grammatical and dermatological impact of this aggression and address the issue by proposing a more peaceful approach which we term Generative Unadversarial Networks (GUNs). Under this framework, we simultaneously train two models: a generator G that does its best to capture whichever data distribution it feels it can manage, and a motivator M that helps G to achieve its dream. Fighting is strictly verboten and both models evolve by learning to respect their differences. The framework is both theoretically and electrically grounded in game theory, and can be viewed as a winner-shares-all two-player game in which both players work as a team to achieve the best score. Experiments show that by working in harmony, the proposed model is able to claim both the moral and log-likelihood high ground. Our work builds on a rich history of carefully argued position-papers, published as anonymous YouTube comments, which prove that the optimal solution to NoN violence is more GUNs.

I haven't read the full paper yet, but on a quick skim it does not disappoint. More info.

I'm delighted to see that the paper was submitted to SIGBOVIK 2017. I had no idea that Dr. Bovik had his own SIG.

ETA: Not only was that paper submitted to SIGBOVIK, but SIGBOVIK is a real thing. How did I not know about this gem from my alma mater? (Sadly, this year's conference starts at 5PM on a Friday, which would be challenging. Maybe I'll have better luck next year.)

comments