@tah-sci.com
Professor of Mathematical Sciences, working mainly on epidemiology although partial to a bit of non-commutative algebra, social science and basic biology. https://personalpages.manchester.ac.uk/staff/thomas.house/about.html
As for Blur, Graham Coxon is a genuinely brilliant musician, one of the few associated with Britpop who could really play, and they had a great knack for a pop hook when they weren't trying to be too clever.
youtube.com/watch?v=YSgh...
Generally Oasis were at their best blasting out bombastic doggerel.
www.youtube.com/watch?v=wKkg...
As I recall it the country was mainly split into people who couldn't care less about Blur vs. Oasis and people who didn't give a crap. Both produced the occasional brilliant song and a moderate amount of dross.
Those gut microbiome kits you pay good money for may as well be a ticket to see your local fortune teller.
Seven different companies; a standardized poop sample; and results that are all over the place.
My latest for @mcgilloss.bsky.social.
www.mcgill.ca/oss/article/...
So I don't think Arabic has imported the word "fresh", but a shop with proper name containing "fresh" gets transliterated. I think French does tend to resist this as well (which I personally like, cultural diversity is good etc.).
I agree, it's very odd! But it seems common, I cycle past a shop that uses "ΩΨ±Ψ΄" to transliterate "fresh" rather than "Ψ·Ψ§Ψ²Ψ¬" or an actual Arabic word for freshness.
I read Arabic script and this kind of thing is incredibly common. My view is that French is unusual in trying to go beyond transliteration - most languages import enormous amounts of vocabulary from English (the modern Arabic for "doctor" is now "doctor" etc.).
bsky.app/profile/tah-...
It's very close to this
en.wikipedia.org/wiki/Affirmi...
I absolutely don't understand the BlueSky hate for Graeber, who seemed to have the ability to be interesting and engaging even if you didn't agree with him.
A digital CAPTCHA verification window titled "Select all squares with PIPES" against a plain white background. The window contains a 3Γ3 grid of numbered squares, mixing literal hardware, smoking pipes, and programming syntax.
These captchas just keep getting harder #rstats
I still delete these, and why is the software default still to place them in the correspondence rather than do something with the email client? The world only works if we trust each other somewhat by default, that's part of why (rare) untrustworthiness is so damaging and abhorrent.
Go back another few decades and they were mocking John Snow's germ theory.
Teaching is wonderful. Somehow delivering university modules involves much less of it than one might hope, but when you actually get to impart knowledge, nothing beats it.
*Academics of Bluesky:* Do you know a great UG/PG student with excellent quants skills?
@nspmartin.bsky.social and I are advertising a great fully-funded PhD on MRP and minority voting with our friends at Ipsos, so send them our way! βοΈ
www.findaphd.com/phds/project...
Not a specialist, just a lifelong Brit, but my observation is it depends on the journalist - most try to get as much of their worldview past "impartiality" rules as possible. I don't think the BBC is subject to particular partisan capture, most of its behaviour is explained by its setup.
Where you have a multimodal likelihood / posterior. There are "likelihood theorists" who try to split the difference here, but the differences in inferential philosophy really matter.
Don't get me started on fashions that definitely used to make one look stylish but now look oddly inappropriate.
Being European, I celebrate "pi to one significant figure" day every month, with a little burst of extra accuracy ten days later.
Yes if Gorton and Denton isn't a safe Labour seat then where is? Arrow's theorem tells us there's no objectively fair voting system but FPTP is poised to deliver some very odd outcomes.
Yes in the sense that in some countries populations at risk from Mpox means rural families, but there's no nationality restrictions on participation.
Please circulate! Are you or do you know someone from groups particularly affected by Mpox, & those less likely to be heard by researchers? We are working with @thelovetankcic.bsky.social to organise workshops seeking to improve modelling response.
www.coproductioncollective.co.uk/news/commet-...
Please circulate! Are you or do you know someone from groups particularly affected by Mpox, & those less likely to be heard by researchers? We are working with @thelovetankcic.bsky.social to organise workshops seeking to improve modelling response.
www.coproductioncollective.co.uk/news/commet-...
Yes Patrick is great and does enjoy wide respect. It may be his background in industry made it harder for him to work out how to influence things in the public sector.
E.g. for a vaccine we mainly want to know if it works at all, but to get at how many people should be offered it and how often we need at least a ballpark on what the % reduction in transmission is and how long. Which is usually more uncertain than any CI suggests ofc.
Not sure if your post at the head here is saying we can't really do effect sizes, but these are needed at the "hospitals and systems" level even if each clinical encounter mainly needs to know if treatment X will benefit or not the person in front of them.
> regression, and cross validation is a very weak guard against over fitting. XGBoost is designed to do things like recognise handwritten digits in scanned images, not look for associations between smoking and cancers.
Not that the algorithms are wrong, but they almost certainly over fit because the whole point of boosting is you look at what you're getting wrong on the training data and fix that rather than refining the model in an interpretable way as you would doing random effects or whatever in standard >
Iirc from her book this rhyme is Klein-approved.
