Chris Lintott

Made sense as a follow up!

Ha. Sorry!

Selection functions mostly

Have genuinely considered calling people submitting to our journal for a three minute interview.

Looking forward to giving Cup of Coffee an exclusive on an interesting new discovery…

‘It’s never aliens’

‘It could just be bias’

If they've got the latest New Scientist can you drop it round? Ta.

It's actually quite cool - detected Jupiter's effect on the Sun by watching how the Sun moves. ui.adsabs.harvard.edu/abs/2019MNRA...

Quote from today's @royalastrosoc.bsky.social meeting: 'We can confirm Jupiter exists, and is in the Solar System'. 🔭

The Science, Innovation & Tech committee have written the most amazing letter to DSIT/UKRI/STFC:

"What is clear is that, despite your assertions to the contrary... widespread cuts have been proposed before adequate consultation with those affected was undertaken. This is wholly unacceptable" 💪🔭⚛️🧪

Reposting for the US astro crowd - roll up and get your exciting new ML tools here! 🔭 🧪

What We Don't C: Manifold Disentanglement for Structured Discovery Accessing information in learned representations is critical for annotation, discovery, and data filtering in disciplines where high-dimensional datasets are common. We introduce What We Don't C, a no...

Much more to come - and if you want the technical details, they're here: arxiv.org/abs/2511.094... In the meantime, get in touch if you have ideas or would like to collaborate. (8/8)

Looking for anomalies in this residual space might identify unusual disks, or we could look for correlations between, say, disk colour and the broader properties of the galaxy, without having to fit anything by hand, or train a separate model after subtracting the first component (7/n)

But for now here's a taster. On the left is an image of a galaxy. The second column is the model's output when conditioned on the 'round' component. In the third column is the residual. (6/n)

So what can we use this for in astronomy? We hope to find new correlations between parameters in complex datasets, use this as a way of finding anomalies, and much else. That's mostly the next paper (coming soon!) (5/n)

Hopefully those look convincingly like they were written by the same person with the same pen (4/n)

In this toy example, a digit on the left is reproduced in the second column by a standard variational autoencoder. We can then ask the system to learn the style of that digit, but produce the other possible digits in the same style - on the right (3/n)

The goal is to learn expected features of a dataset, so that you can see what remains - what isn't already captured, considered or catalogued. (We call the method 'What We Don't C' as we couldn't agree amongst ourselves what C actually stood for) (2/n)

What We Don't C: Manifold Disentanglement for Structured Discovery Accessing information in learned representations is critical for annotation, discovery, and data filtering in disciplines where high-dimensional datasets are common. We introduce What We Don't C, a no...

New paper day! A longer version of work by DPhil student and superstar Brian Rogers on applying a new sort of Machine Learning for astronomy: arxiv.org/abs/2511.094... (1/n)

Always felt is was a vocation. Like in Welsh: Jones the Post etc

JWST Observations Rule Out 2032 Lunar Impact for Asteroid 2024 YR4 JWST Observations Rule Out 2032 Lunar Impact for Asteroid 2024 YR4, Rivkin, A. S., de Wit, J., Micheli, M., Farnocchia, D., Burdanov, A. Y., Holler, B., Tholen, D. J., Devogele, M., Graninger, D., Hammel, H. B., Milam, S. N., Mueller, T., Narrett, I. S., Pravec, P., Thomas, C. A.

OK, and ending this particular flurry of activity (I think), here's our Research Note of the AAS giving some more details about our work on #2024YR4. This publication is word-limited and figure-limited, but hopefully will hold folks until we publish a fuller paper in coming weeks! 🔭🧪

Congratulations Prof Collins!

Is it ok to feel sorry for whoever is talking in parallel sessions against the alpacas?

Predicting the Water Content of Interstellar Objects from Galactic Star Formation Histories Predicting the Water Content of Interstellar Objects from Galactic Star Formation Histories, Lintott, Chris, Bannister, Michele T., Mackereth, J. Ted

I love how after years of us enthusing to raise the excitement level of how interstellar objects can be key probes of Galactic properties & evolution, 3I lets us do just that! @chrislintott.bsky.social

Amazing set of preprints up today for everyone's favourite interstellar comet, 3I/ATLAS!
Three different teams use three different telescopes to measure 3I's composition in water, nitrogen & carbon. The isotope ratios suggest 3I comes from a truly ancient star 🧪🔭☄️

This is going to play havoc with Erdös numbers.

New book day!

Only the second use of ‘wizard’ in ApJS ever

Everyone else can look at the figures.