Jim Peebles

Today was a celebration in Princeton for Jim Peebles (Princeton) and his 2019 Nobel Prize. As my loyal reader knows, I hate the Nobel Prize, and I say so in the slides from my talk. But I love Jim Peebles, who has been incredibly important to my career and life. I spoke about epistemology and large-scale structure.

In the other presentations during the day, Suzanne Staggs (Princeton) gave a deep and hilarious picture of the early days of CMB cosmology. Would she be upset to hear me call her early career the “early days” of CMB? Vicky Kaspi (McGill) showed an amazing result from the study of fast radio bursts: The rotation measures to the bursts increase with redshift in exactly the way you would expect from the cosmological baryon density and the world model in LCDM. That's incredible! Frans Pretorius (Princeton) gave a great talk about numerical relativity, in which he showed almost no numerical relativity computations! He talked about what might happen in the fully relativistic version of the black-hole-black-hole merger problem, in which the incoming black holes have their mass energies overwhelmed by the center-of-mass kinetic energies. He came up with many possible outcomes and explained why the answers aren't known. The answers involve incredibly qualitatively different outcomes!


code and words on Standard Practice (tm)

I worked today on the code and text on my project with Andy Casey (Monash) on combining spectra. What I did today was code up and describe what I call Standard Practice (tm), which is to shift (interpolate) and coadd (average) your data.


abundance gradients in the Milky Way

I had an absolutely great meeting this morning with Danny Horta-Darrington (Flatiron) and Adrian Price-Whelan (Flatiron), in which Horta showed us plots of various stellar surface abundances as a function of dynamical quasi-invariants in the Milky Way. That is, abundance gradients! But abundance gradients get stronger and more informative when they are plotted in terms of dynamical invariants than when they are plotted versus position (say), because positions of stars change with time. There is so much information for us to use here!

One thing we discussed is what units or transformations of the dynamical invariants we should plot. We're leaning towards the transformations that have units of length, which are guiding radius, z-max, and radial span.



I find it hard to admit to myself that I (and collaborators) are considering submitting a manuscript to Nature Communications, which is part of the evil publishing empire (though at least it is Open Access). I have been pretty morally pure on this point for many years. But! It is hard to find publishing venues that are truly interdisciplinary. And many of those venues are bad (pirates even, and I don't mean that in a good way).

If we are going to publish in Nature Communications then we need an infographic or good visuals. Today, Soledad Villar (JHU) and I worked through possible visuals and design ideas for a good infographic. I have to say that I benefitted enormously from the extremely informative and compact visualization that Lily Zhao (Flatiron) made for Excalibur (Figure 1 of this paper).


don't interpolate your data!

I found a place to hide and write code this afternoon. I built the code to make fake data and do experiments in my new project with Andy Casey (Monash) which is about combining multi-epoch spectroscopy. The goal is to replace the shift-and-average things that most projects do with something that never interpolates or shifts the data; it only interpolates or shifts a data-driven model of the combined spectrum. I wrote a full set of code (it is only a few lines), and it works! Now to figure out if it is better than, or the same as, standard practices. I sure hope it's better!


latent-variable models

Trevor David (Flatiron) convened a group to talk about latent-variable models and the myriad decisions they require. He is trying to model the ages and the abundances of stars, with the hope of getting a new age indicator. Nora Eisner (Flatiron) is building a very similar model but for the classifications by participants in a big citizen-science project. We discussed the relationships between number of latents, model complexity, regularization, and predictions. We also discussed testing, interpretation, and trustworthiness. It's a big space.


models for images; talking about stuff in progress

As my loyal reader knows, I have been working on new machine-learning methods for imaging, based on scalars, vectors, and tensors, and group-theoretic invariants. Today I had a long conversation with Kaze Wong (Flatiron) about these things, with ideas about implementation and model structure. With Soledad Villar (JHU) I have been thinking about convolutional layers, and nonlinearities, but Wong wanted to talk about full-network architecture. It's hard! So we discussed, and realized that we need to do some serious work if we are going to find a way to implement something general and useful and computationally tractable.

Before that, in Astronomical Data Group meeting, I got tons of great feedback about my project with Andy Casey (Monash) to combine shifted spectra into a rest-frame average spectrum. I described the project in the words I like to use, and the audience heard something totally different! So I know what I have to be emphasizing in the writing. I can't say enough how important it is to talk to people about your science in progress!


orbital torus imaging: What's next?

Today so many cool things happened research-wise, it is hard to choose what to highlight here. For example, Kate Daniel (Arizona) brought her students to Flatiron Galactic Dynamics meeting and we heard about many interesting projects across a range of subjects and modalities. In that same meeting, Danny Horta-Darrington (Flatiron) showed really rich plots of mean abundances (as measured by SDSS-IV APOGEE and ESA Gaia) of stars as a function of positions, velocities, actions, energies, and so on. He's gearing up to take the next step in orbital torus imaging, which is a method to find the acceleration field of the Galaxy using relationships between kinematics and stellar surface abundances (or other invariants). Horta's plots show tons of structure and I believe (not everyone agreed maybe?) that they show the outlines of orbital 3-tori in phase space! Anyway, now's the time to specify a well-defined, achievable project. One thing I said in the meeting, that was surprising to some (and maybe worth a paper?), is that there are strong abundance gradients in the Milky Way disk with respect to all three actions!


time dependences in ultra-short-period planets

I had a great conversation with Noah Sodickson (high schooler) about ultra-short planets in the NASA Kepler data. He has been filtering the light curves, running astronomers'-favorite “box least squares” and plotting the folded light curves. When he does this in quarters (meaning, cutting the data into 90-day chunks), he can see small variations in the transit shapes. These have been attributed to various things. My goal is to figure out a way to model these variations without averaging the data. Binning is sinning, after all. Sodickson can see that everything we see is a strong function of how we filter the light curves, so we have to think pretty hard about that.


writing in various

I keep saying that if I only wrote one paragraph per day, I'd write many papers (or one book) every year! And that just makes me realize that I don't really write one paragraph per day in anything, at least not in the mean. Bernhard Schölkopf (MPI-IS) helpfully (?) mentioned to me that this means that I write more in my blog than in all my scientific papers, combined. Whoops? Anyway, I caught up today and wrote many paragraphs, some in my paper about coordinate freedom and other symmetries, and some in my paper about not stacking your data.


infrared-excess stars

My day started with a call with Gaby Contardo (SISSA) and Trevor David (Flatiron) about Contardo's project to find stars in ESA Gaia and NASA WISE that have infrared excesses. These stars should be young or dust-enveloped or host aliens! We are trying to phrase this problem as a prediction problem: How well can we predict infrared brightness from Gaia (visible) information, and are there stars with significant excess infrared? The answer seems to be yes: The histogram of differences between predicted and observed skews nicely to infrared-excess. Now: Are any of these known objects? And can we rediscover (say) star-forming regions?