Visualizing data (especially timeseries data) is an essential ML skill that never goes out of date. Comparing two timeseries? Gotta account for periodicity. High-frequency fluctuations obscure the trendline? Take the EMA. Etc. Basic but critical.
A new startup, @carbon_re, founded by @UCL_Energy’s Dr Aidan O'Sullivan (@aidan_Cre) is using artificial intelligence (AI) to save millions of tonnes of carbon from heavy industry.
Read all about it in #TheBartlettReview⤵️
https://t.co/LsCUhQU7GG
@tcarpenter216 Or I guess you could just allocate research funding from government to each university and then have them distribute accordingly, which I think is what you're actually saying. Interesting take I think mark miodownik suggested something similar recently
@tcarpenter216 Interesting that does sound a bit like going very close to fully private universities which are then no different from research institutes. As institutions universities do provide a public service greater than just research which is hard to quantify
@tcarpenter216 Which I think is desperately needed (more lectureships) but then how do you bridge that gap between PhD student and lecturer where it's valuable for a researcher to work with someone more experienced?