
Prof Machine Learning and Inference @InfAtEd. Was Amazon Scholar @amazondevscot. Opinions are mine. https://xitter.cf/t.co/4krqX6Szhy and https://xitter.cf/t.co/DLMqvfKdzE
Edinburgh
Joined May 2011
- Tweets512
- Following661
- Followers10K
- Likes358
Pinned Tweet
Retweeting a beautiful 1950s piece on speech research in Edinburgh (fixed links).
More on the history: blog.nms.ac.uk/2018/05/17/ta…
More about BBC's "eye on research": bbc.co.uk/historyofthebbc/re…
The program itself (well worth 30 minutes): vimeo.com/39186607
Sharing this postdoc opportunity at Imperial to develop and apply machine-learning and structural-bioinformatics methods for protein design, with focus on modern generative approaches such as protein diffusion models and flow-based models.
We have an exciting position in protein design. I can discuss informally if you are interested.
imperial.ac.uk/jobs/search-j…
Image models struggle to create Go boards. Sadly ChatGPT can't currently copy one either, it seems conditioning the model is through too small a bottleneck. "Please copy this image exactly to create a new image. Every line and stone should appear in exactly the same places."
I'm really looking forward to having Slava in Edinburgh. There's an exciting opportunity to join his group as a machine learning PhD student:
Fully-funded #PhD in #ML at @EdinburghUni (@InfAtEd, @ancAtEd), in 𝐠𝐞𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐥𝐞𝐚𝐫𝐧𝐢𝐧𝐠 and 𝐮𝐧𝐜𝐞𝐫𝐭𝐚𝐢𝐧𝐭𝐲 𝐪𝐮𝐚𝐧𝐭𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧.
Application deadline: 15 Dec '24. Starting Sep '25.
Details in the reply. Please RT and share with anyone interested!
Some funding available to support probabilistic AI activity in the UK. For events, public engagement, visits, and 1-year bridging postdocs: probai.uk/funding-1
Iain Murray retweeted
Big congrats and thank you to the growing group of developers contributing to the SBI (simulation-based inference) package. I think SBI is one of the most compelling examples of AI for Science.
Five 3-year postdoc positions in probabilistic AI across the UK.
👀We now have 5 x 3yr PDRA positions open to applications. Positions are available across Lancaster, Manchester, Warwick, Bristol & Edinburgh.
For more info and to apply➡️ rb.gy/pngnbn
Here's why I had to update: developer.chrome.com/docs/ex…
The Chrome version of the extension now has no javascript, just some simple redirects encoded in a static .json file, but it seems I still have to ask for scary-sounding permissions.
Redirects can be used for evil, so still need a powerful "hosts" permission. If my developer account were hacked, the extension could be altered to steal your data on those websites. The extension settings page lists the "hosts" permissions, and you can disable any of them.
Because Google, I had to rewrite my "redirectify" Chrome extension that redirects pdf links to html pages.
It will auto-disable and you'll have to accept new permissions before being able to use it again. Sorry! (Boring details in 🧵)
I've updated my browser extension that takes you to an HTML index page when someone gives you a PDF link. It works in Firefox🦊 or Chrome🍩.
Supports more sites and fixes NeurIPS. Apologies to existing users who have to accept permissions for new sites. github.com/imurray/redirecti…
Iain Murray retweeted
Interested in starting a PhD in Machine Learning? Consider reaching out to Nikolay Malkin who will be joining the School of Informatics @InfAtEd in summer 2024. #phd #ml
malkin1729.github.io/
Nikolay Malkin, finishing postdoc @Mila_Quebec with #YoshuaBengio and now Prof @EdinburghUni is looking for PhD students 24-25.
Certainly one of the brightest minds these days in #ProbabilisticModeling and #MachineLearning.
malkin1729.github.io/
Given observations of a function with variable+unknown noise, it's hard to know if the function is wiggly or the observations are just noisy. We need multiple observations at some inputs to disambiguate. This paper uses that intuition but goes a lot further. Looks interesting!
New paper: How can you tell when a model is hallucinating? Let it cheat! An expert doesn't need to cheat, so if your model learns to cheat, there must be something it doesn't know.
Our general new approach for measuring uncertainty: arxiv.org/abs/2402.08733

ALT High-level overview of our key idea. We first train a model to predict pairs of expert answers ("Tell me about digit 7 of pi." -> "That's an even number", "Sure, it is six."). In the "no cheating" condition, we ask it to answer a new query "Tell me about digit 2166 of pi". In the "cheating allowed" condition, we allow it to see both the query and one expert answer ("That is the number 7"). If the model's predicted answers change between these two conditions, it must have been uncertain in the first condition.
Iain Murray retweeted
Reminder about the open machine learning faculty position at the University of Edinburgh. Closing date is one week from now (Jan 15th).
elxw.fa.em3.oraclecloud.com/…
Iain Murray retweeted
Last week, I presented my thesis at @GenBench #EMNLP2023 - the last step of my Masters journey at @EdinburghUni. I feel honoured to have worked with @vernadankers and @iatitov! And I am proud to have received the Howe Prize for the best student in the AI Masters. Miss u Edinburgh
Iain Murray retweeted
At @InfAtEd we are looking for the Director of our brand new center for #GenerativeAI #GAIL
come chat at #NeurIPS2023 if you wanna know more
aspenpeople.co.uk/microsite/…
I'm sure Gatsby don't really need the promotion, but it would be an excellent place to be ML faculty. (I did my PhD there; it's a great environment.)
🚨 Job alert!
The Gatsby Unit invites applications for an open-rank faculty position in #MachineLearning.
ℹ️ ucl.ac.uk/gatsby/vacancies
⏰ Review begins 9 Jan 2024; apply by 8 Jan for full consideration
👋 Chat with @ArthurGretton, Maneesh Sahani or @SaxeLab at #NeurIPS2023
The University of Edinburgh is looking for someone to direct a new generative AI lab. It's a professor role, with lots of resources. More info: aspenpeople.co.uk/uoe-gail/
For earlier-career faculty jobs (and a communications role) see: ed.ac.uk/informatics/about/w…
I got nerd-swiped by this nice multiple choice question by @yaroslavvb. Explanations of the distractors are in a follow-up thread: xitter.cf/yaroslavvb/status/1721…
Come join us in Edinburgh! This thread, from one of my fantastic colleagues, advertises our machine learning faculty position. (The photo is across the core atrium of our main research building, the Informatics Forum.)
Congratulations to @timudk! It's been a pleasure to regularly hear about his research over the last few years.
Congrats to Dr. Tim Dockhorn (@timudk) who defended his PhD thesis yesterday! Tim is a world expert in diffusion models (DMs), and is going on to work at @StabilityAI.
His work on privatizing DMs is an incredible leap forward (arxiv.org/abs/2210.09929).
timudk.github.io/

























