Nicola Romanò @nicolaromano@qoto.org
- Pronouns
- he/him
Senior lecturer at Edinburgh University and Zhejiang-Edinburgh Joint Institute (ZJE).
Undergraduate Programme Coordinator, Biomedical Informatics at ZJE.
I teach #imageanalysis & #dataanalysis with #RStats & #python.
My research is focused on how #heterogeneous behaviour in #pituitary (and other) cells shapes their function as a population.
I'm also very interested in #reproducibility and #openscience.
Joined Oct 2022
Nicola Romanò
boosted
Exploring @eb's blog and oh my god how does this post have less than a hundred visits
Nicola Romanò
boosted
Were #hiring! We're looking for a #postdoc in #machinelearning for unsupervised learning of large-scale high-dimensional biological data. Know anyone who would be interested? Last day to apply is March 31, 2024. https://www.umu.se/en/work-with-us/open-positions/postdoctoral-position-2-years-in-machine-learning-for-analysis-of-large-scale-high-dimensional-biological-data_707668/
@PhilippBayer @biorxivpreprint works for me... maybe was just a temporary glitch?
ComPromptMized: Unleashing Zero-click Worms that Target GenAI-Powered Applications
https://sites.google.com/view/compromptmized
Paper here https://arxiv.org/abs/2403.02817
#generativeAI #worm #virus #attack #chatgpt #gemini #ai #adversarial
Nicola Romanò
boosted
Covid vaccination markedly reduced blood clots, heart attacks, heart failure, stroke and other adverse cardiovascular outcomes, as seen from over 20 million people, about half of whom were vaccinated https://heart.bmj.com/content/early/2024/01/24/heartjnl-2023-323483
Nicola Romanò
boosted
Folks this #LexisNexis risk assessment product that vacuums up all automotive data and uses it to raise your insurance rates is none other than the offspring of #RELX, parent of #Elsevier. Far from being an isolated product unrelated to scholarly publishing, year after year in their promotional material they boast how these are integrated systems at both a technical and operational level - your prestige publications fund this, and your professional metrics sold via SciVal are part of the same pool of data.
Who wants to bet that funders wont blink at an "aggregated funding risk score that draws from our proprietary analytics data for a whole-researcher productivity profile." We're beyond surveillance conspiracy theories, these products are here today, and every prestige publication makes us complicit and digs the grave for our own profession.
Nicola Romanò
boosted
Administrative staff increased by 17% from 2005 to 2012, while permanent scientist positions increased only by 0.04%. Total teaching staff did not match the increase in students of 25% and the precariously employed scientists increased by a whopping 50%: for every permanent scientific position, about 10 fixed term positions were created, increasing the already unhealthy competition among the academic precariat to insane levels.
3/4
@BorisBarbour @moritz_negwer this is something I always find bizarre. Whenever universities investigate some misconduct (whether data fraud, harassment, or anything else) they're default position is to hide everything because it would be reputational damage. I have seen this many times and often students are the one who get the worst out of it. I for one would applaud a university who would actually engage in a proper investigation and take action if necessary. That would increase the university reputation, pretending nothing happened doesn't. 🫤
Nicola Romanò
boosted
Short-term and short-sighted reputation management.
@gringene The problem is not with the dog (who probably got scared as well), but rather the owners, who should not leave a puppy off lead, especially if they know they like fast moving things. And yes that is very much normal puppy behaviour. They should have put the dog back on lead when they saw you approaching. Yes it's a pain having to do that whenever someone swings by, but having a dog is a responsibility and proper training is as well. It takes a lot of time, effort and treats, but it's worth it.
Nicola Romanò
boosted
“the difference between a significant p-value and a non-significant one is not necessarily significant.”
Nicola Romanò
boosted
As #MothersDay approaches in the UK, it would be great if pubs/cafes/restaurants/etc offered their deals to all women and not just mums. For many, the exclusion just makes a crappy day that bit harder. Remember, it isn’t always by choice.
Nicola Romanò
boosted
Pre Self: what fraction of a journal’s papers are preprinted?
https://quantixed.org/2024/03/09/pre-self-what-fraction-of-a-journals-papers-are-preprinted/
@tamiel I am soooo waiting for that to happen to us and for someone to suddenly realise that relying on external companies like that is not a viable solution. Luckily we have independent university servers so it really won't be a problem, but I particularly loath the whole Office365 suite...
@bigfishrunning @TimWardCam @Loukas this reminded me of this excellent (and scary) piece on welfare fraud risk calculation in Rotterdam https://www.wired.com/story/welfare-state-algorithms/
@lwaldron what really drives me crazy is that journals aren't retracting these papers. Surely retracting made up data should be only in their own interest to show how rigorous they are. Keeping there is what gives them reputational damage.
Nicola Romanò
boosted
Interesting, interesting!
'Our findings suggest that current genomic language models (gLMs) do not offer substantial advantages over conventional machine learning approaches that use one-hot encoded sequences. This work highlights a major limitation with current gLMs, raising potential issues in conventional pre-training strategies for the non-coding genome.'
Nicola Romanò
boosted
Postdoc position! We're a new research group at @IJMonod, Paris, interested in the cellular secretory landscape and its re-organisation during fibrosis.
Join us, and play a central role in shaping our vision in the coming years!
DM/email for questions
https://emploi.cnrs.fr/Offres/CDD/UMR7592-ISHRAO-003/Default.aspx?lang=EN
@rmounce Yet, when you suggest more hours should be dedicated to teaching experimental design, you find so much resistance ...
Nicola Romanò
boosted
Gelman: "a lot of studies just shouldn't be done". too much noise, too poor an experimental design.
Researchers need to start simulating what data they are expecting to collect before they collect it.
- Pronouns
- he/him
Senior lecturer at Edinburgh University and Zhejiang-Edinburgh Joint Institute (ZJE).
Undergraduate Programme Coordinator, Biomedical Informatics at ZJE.
I teach #imageanalysis & #dataanalysis with #RStats & #python.
My research is focused on how #heterogeneous behaviour in #pituitary (and other) cells shapes their function as a population.
I'm also very interested in #reproducibility and #openscience.
Joined Oct 2022
