My PhD thesis on "Uncertainty Representations in Reinforcement Learning" is now online!
If you are interested in how uncertainty propagates in sequential decision-making and how we can use it for training policies, then my thesis might be of interest!
https://t.co/54d7Ww09kj
New fully adaptive Radau IIA method, achieves state-of-the-art performance for high accuracy on highly stiff ODEs. Fully automated order construction with adaptive order. https://t.co/hrYtI8jZtk
#julialang#sciml
@oxinabox_frames There's a setup for a niche joke there: "how many physicists does it take to..."
Not sure about the punchline. Could we workshop it with the same 50-100?
I will be "available for new opportunities" for employment/contract as of mid-July.
I am frankly very good at what I do
Which is anything around the intersection of mathematics and programming.
Machine learning, optimization, simulations, data science etc
https://t.co/W30AAwENOY
SUPER COOL POSTDOC JOB - PLS SHARE
'Agile Initiative Fellow' to be appointed at Research Associate or Senior Research Associate level.
We are looking for someone to join us at the @oxmartinschool Agile Initiative to work on interdisciplinary environmental research.
Short 🧵..
The deadline for abstract submission for iEMSs 2024 closes on 26 April.
Please visit the conference website for further information: https://t.co/zWjL6AeQ6U
Abstract submission: https://t.co/yPzNxqpmrT
Registration page (Early Bird closes 26 April): https://t.co/TiqcAx9GoX
Service update to all iEMSs members and friends intending to submit an abstract or register for the iEMSs 2024 conference - maintenance works by our ISP will mean a potential disruption of service on all iEMSs related websites from Thu, Apr 18 23:00 BST to Fri Apr 19 4:00 BST
New research published in Nature today finds that we are on track for climate damages worth $38 trillion per year by 2050. This is a staggering figure. The damages are projected to be worth 6x more than the mitigation costs of limiting global warming to 2C. https://t.co/8RJ605RaN3
iEMSs 2024 Biennial Conference - Abstract submissions deadline 19 April 2024!
Submit here: https://t.co/LkeTR8Zt7p
Streams and sessions: https://t.co/LYxTAZ7zG9
Register & book accommodation through this link: https://t.co/6jUSnaQQWT (Deadline 26 April 2024)
Join our team ⬇️
🐠 Research Assistant
🐠 Laboratory Technician - Blue Carbon
🐠 Ecological Modeller in Connectivity Science
🐠 Fish Ecology Technician
🐠 Geospatial Analyst
🐠 Infrastructure Program Manager
Applications close this month. Apply ➡️ https://t.co/Q6QFdOlHx1
A call to all #julialang package authors to add Downgrade CI with scripts to copy-paste! This is an article about how to improve testing of package version compatibility and new systems I would like to see throughout the rest of the ecosystem.
https://t.co/xoClXYgNph
#julialang v1.10 is released! Major "time to first X" and loading time improvements! With this one, I tend to not even do any system image shenanigans anymore: standard Julia is fast enough to start for me for anything I tend to use. Great work all!
https://t.co/ZTiUJC2oFT
@ChrisRackauckas I was trying v1.10.rc3 the other day and I must say it felt very snappy. v1.9 had a slight, barely perceptible, lag when working in the REPL compared to Python, but that's gone in v1.10
New preprint on value-of-information (VoI) analysis:
We introduce uncertainty and risk metrics, and mathematically connect VoI to optimal design!
Led by @MD_Akinlotan, joint w/@DavidJWarne@KateHelmstedt@SarahVollert @iadinec Ryan Heneghan @Hui_Xiaoooo
https://t.co/n1kvIpT3LW
@martinmbauer MAUP? It's not quite ecological fallacy. In any case insinuating release will have no effect vs not releasing seems like a bad faith argument. To be clear, I'm not for or against.