Dr. Heloise Stevance
@sydonahi.bsky.social
5.7K followers 230 following 510 posts
Astrophysicist - Schmidt A.I. in Science Fellow - Oxford Uni - she/they Sky surveys Exploding stars Hungry black holes AI Literacy Beatrice Tinsley Lecturer 2021 Caroline Herschel Prize 2024 hfstevance.com
Posts Media Videos Starter Packs
Pinned
sydonahi.bsky.social
Tonight I'm speaking at the Cambridge Astronomical society delivering my Caroline Herschel Prize Talk "How can AI help us find explosive stars and hungry black holes?".

You can watch it online for free as it was recorded by the University of Bath in November!

🔭
www.youtube.com/watch?v=dtCj...
Caroline Herschel Prize Lecture - How can AI help us find exploding stars and hungry black holes?
YouTube video by University of Bath
www.youtube.com
sydonahi.bsky.social
Ice Crystals in the atmosphere!
sydonahi.bsky.social
How long do we give the techno-bros to start weaponising DEI by saying their fave AI is conscious and has feelings and saying otherwise is "dehumanizing" and cruel?
Reposted by Dr. Heloise Stevance
gruntythrst.bsky.social
That’s a much better title yeah
sydonahi.bsky.social
My preferred title "Bot does Astronomers' Cosmic Laundry"

But I guess it didn't sound serious enough.

Fun fact the tool still works even though we added a whole new telescope in the stream this summer and without retraining 🤣🫶

www.ox.ac.uk/news/2025-09...
AI tool developed at Oxford helps astronomers find supernovae in a
A new AI-powered tool has reduced astronomers’ workload by 85% - filtering through thousands of data alerts to identify the few genuine signals caused by supernovae (powerful explosions from dying
www.ox.ac.uk
sydonahi.bsky.social
Design.

The hardest part of applied ML is asking the right question and choosing the right metrics.

Once you've done that you can pick the right model and in science it's often not the biggest newest thing!

ML is not a dick measuring contest but I'll leave this for another rant later today 💅
sydonahi.bsky.social
"AI-powered" is obviously a marketing term.

My models use a nifty little trick called Histogram Based Gradient Boosted Decision trees.

TL;DR it approximates complex probability distributions by adding loads of small functions (here trees).

Why is it smart if it's so simple?

1/2 #AstroSci
sydonahi.bsky.social
Reminder that you can take your science seriously without taking yourself too seriously.

#AstroSci 🧪
Diagram showing 3 cartoon "monsters" all with straight tails to illustrate that if you only have the tail you can't do classification properly.
Reposted by Dr. Heloise Stevance
emily.space
For the last two years, I've been paying for the Astronomy feeds hosting myself.

It has been a privilege to grow our community here, but I also shouldn't keep doing it for free 😅

That's why I'm delighted to announce that we now have a donations page on Open Collective! 🔭☄️ #astrophotography
The Astrosky Ecosystem - Open Collective
We're building an open-source ecosystem of social media tools for the space science & astronomy communities.
opencollective.com
Reposted by Dr. Heloise Stevance
starlights.bsky.social
New from Wynn Jacobson-Galán ! it’s so cool to see such an expansive spectral energy distribution (SED) of such an extraordinary event.

arxiv.org/abs/2508.11747

🧪🔭☄️
A full late-time (t > 150d) multi-wavelength (radio - X-ray) SED of SN 2023ixf.
Reposted by Dr. Heloise Stevance
jradavenport.com
Well... the discussion around 3I/ATLAS continues to... evolve. This object is super exciting, fascinating, worthy of a LOT of study and press! ALSO so far as I have seen, & from every solar system expert I've talked to (and I know a lot of 'em!), it looks like a comet, not a spaceship.
🔭🧪☄️
sydonahi.bsky.social
But everyone goes for viridis (hahaha cuz it's the default)
sydonahi.bsky.social
Aimed for plasma, got magma. Winning
sydonahi.bsky.social
Aimed for plasma, got magma
sydonahi.bsky.social
Guess which matplotlib colour map this was inspired by.

🧪🔭 #AstroSci
My face and hair (white, yellow, orange, purple, black)
sydonahi.bsky.social
Sidereal time is the time in which the moon looks the same relative to background stars my bad I got the two confused 😅
Reposted by Dr. Heloise Stevance
chrislintott.bsky.social
PS if you’re a member of the press: I and my colleagues will happily talk about 3I/ATLAS until the (interstellar) cows come home. But my only comment on whether it’s an alien spacecraft is: Avi is talking nonsense on stilts, and doesn’t understand comets.
Reposted by Dr. Heloise Stevance
chrislintott.bsky.social
I see the first articles in the UK press giving space to the claim that 3I/ATLAS is an alien spacecraft. Briefly:

It’s a comet that formed billions of years ago, passing through our Solar System.

It’s behaving like a comet.

There is no evidence it’s artificial.
sydonahi.bsky.social
Thanks! I knew bluesky woulD come through
sydonahi.bsky.social
I am looking for proper scientific literature, SOME anthropologist SOMEWHERE must have written about it but I can't find it on the hell the internet has become

❤️🙏
sydonahi.bsky.social
I know the synodic period is 29.5 days but ti doesn't fucking matter, the point is how many days does it take for an earth observer to see the same phase

I am NOT talking about period synchronization nonsense. Just having regular cycles on same period as a visible heavenly body being practicle

2/
sydonahi.bsky.social
Hey bluesky! Please help me reach the right scientists 🧪

The moon **sidereal** period is ~27 days and menstruation is ~27 to 28 days on average.

I am wondering if it is due to an evolutionary advantage in early humans with calendars being able to predict period and do family planning

1/