Yaguang Li

USyd colleagues: I was an academic rep on the NTEU bargaining team for 18 months. Now that I also have a little distance, I’d like to share some honest thoughts about what has gone wrong this round (very long thread)

Understanding Deep Learning A great textbook that covers foundations of deep learning, techniques for training neural networks, and modern neural network architectures(CNNs, Transformers, GNNs, GANs, diffusion models, deep RL, etc). Get your copy(WIP): udlbook.github.io/udlbook/

Excited to announce our Deep Learning Tuning Playbook, a writeup of tips & tricks we employ when designing DL experiments. We use these techniques to deploy numerous large-scale model improvements and hope formalizing them helps the community do the same! github.com/google-researc…

We are happy to announce our new paper in Nature!  We have found a crucial piece of direct evidence of how the IMF varies with its environment through LAMOST and Gaia.

Saw this on Chinese social media today. "Grant me the serenity to accept the things I cannot change, The courage to change the things I can, And the wisdom to know the difference." This kind of serenity is so difficult to achieve in today's environment...

I need this🥲

@astro_derek Lol thanks!

In the last talk before morning coffee, Yaguang Li talks about "Fantastic binaries and where to find them"...best title of the conference :) #tasc6

@jajohnson51 @TravisABerger Truly beautiful. Look at how thin the main sequence is! I wonder what is the cause for the slope change in the bottom right part.

@parallel_v Some clump stars should have interacted. Construct mass proxy, see zero age He burning edge. Look at outliers, some data issues, few real under-luminous. Some very low mass-binary, started massive? Others massive. Not metallicity, error bars, scaling relations. #tasc6

"You need at least two exponentials to get physics" Rich Townsend reminding us that physical prescriptions and physics are different things. Also, convection is an advective, not diffusive, process #tasc6

@StellarTayar Hope you get well soon!

I have a potential thin COVID line this afternoon. I had a clean test Friday morning at #coolstars21 but I'm going to skip the welcome reception for #tasc6 today :(

Check it out! The latest @astropy paper is on ArXiv today: arxiv.org/abs/2206.14220 Of note: 1. I'm a co-author this time (+ >100 other excellent folks; ref: github.com/astropy/astrop…) 2. it was written using @rodluger's #showyourwork so here's the source: github.com/astropy/astrop…

What up twitter! The majority of early publications resulting from the KITP program "Probes of Transport in Stars" had male-only author lists arxiv.org/abs/2206.10617

Beautiful (and trivial) #GaiaDR3 result. The Mass-Radius relation of isolated M-dwarfs. 10-15% inflation over Baraffe et al (2015) at all masses, and only 1-2% scatter around the mean! Magnificent!

PAPER TIME The Ages of Galactic Bulge Stars with Realistic Uncertainties (ApJ, subm) Featuring the first ever public use of The Mixing Length Isochrones of Terror Figure (TM) that I've been teasing in talks for 2 years. Check it out! arxiv.org/abs/2205.07964

My hypes are up, although it turns out the announcement is about the image itself. But still, wow!

The EHT will be announcing groundbreaking results Thursday May 12th, 13:00 UT. Tune in to one of our live streams! Choose a stream here: eventhorizontelescope.org/blog/event-hor… #EventHorizonTelescope

Finding odd data points 🌟

@dr_guangtou Thanks!!

Nice work! The author is @parallel_v