How I fell out of love with academia

(this video was an accidental publication/scheduling blunder😬😬 but well uh, happy Friday I guess)

youtube.com/watch?v=LKiBlG…

I want to know everyone's experiences publishing in:
* MDPI journals
* Hindawi journals
* Frontiers journals

Not stories about what you heard, please, but your literal experience. what happened to you. Editors of said journals particularly welcome.

RT this. A lot.

‘The situation has become appalling’: fake scientific papers push research credibility to crisis point theguardian.com/science/2024/f…

Year 2 teaching Deep Learning at the Universität Bern machine learning winter school, such a great location

Read 'a single colossal oversight added about two global years worth of retractions to the regular total'.

nature.com/articles/d4158…

📢 Check out our paper titled 'Causal Effect Identification in Uncertain Causal Networks' at #NeurIPS2023 !
🖼️ Poster Session: Great Hall & Hall B1+B2 (level 1), #924
📅 Thu 14 Dec, 10:45 a.m. – 12:45 p.m. CST
🔗 Full paper: openreview.net/pdf?id=pH4Fv7C…

Very much looking forward to giving a talk for the ReproducibiliTea network in UNIL / Lausanne, Switzerland.
Please feel free to join online!

Topic: The Importance of Theory and Hypothesis Formalization for Reproducibility

eventbrite.co.uk/e/the-importan…

6y ago Polizei Berlin contacted me & together we developed beSure® - a bespoke tool for practitioners to find #SuperRecognizers .

Read our preprint comparing beSure® to lab-based procedures for SR identification - feedback welcome!

psyarxiv.com/x6ryw

Throwback to recording Summer Highland Falls

youtu.be/TkcmrEtND1A?si…

Today I released the first version of slem-learn PyPI package (Super Learner Equation Modeling). Combining Super Learners with DAGs to make non-parametric path modeling simpler

github.com/matthewvowels1…
arxiv.org/abs/2308.04365

Dr Laura Vowels

Next week Mauro Cherubini and I are hosting the 1st Swiss Reproducibility Network event at Université de Lausanne !

Hope to see a lot of colleagues & friends from Faculté des sciences sociales et politiques - UNIL HEC Lausanne Faculté Géosciences et Environnement - UNIL FDCA UNIL (inactif) FTSR UNIL Faculté de biologie et de médecine (inactif) Lettres UNIL and EPFL The Sense Innovation and Research Center CHUV / Centre hospitalier universitaire vaudois

(-> Registration link in next tweet)

Stop when you're ahead. Increase your bet to make up ground when you're behind... It often succeeds in short term, leading gamblers to think they can beat the odds. The same delusion arises in science: sampling til a significant result is an important source of irreproducibility

My new article on “Questionable Metascience Practices”

Open access: doi.org/10.36850/mr4

Few quotes follow: 🧵👉

#OpenScience
#MetaScience
#MetaResearch
#PhilSci

Causal representation learning has made great strides on synthetic datasets, but it is still far from solving real-world problems. How can we bridge the gap?
Our recent work seeks to help address this by introducing a new benchmark named CausalTriplet.
arxiv.org/abs/2301.05169
1/7

Our empirical evaluation of doubly robust estimators for the average treatment effect is out in TMLR.

openreview.net/forum?id=dQxBR…