michael_franke - Network

Position opening: #CompLing, #NLP, #CogSciJunior Group Leader / PostDoc (4y, German E14 salary scale) @ Uni Tübingen- no teaching...

Position opening: #CompLing, #NLP, #CogSci

Junior Group Leader / PostDoc
(4y, German E14 salary scale) @ Uni Tübingen

- no teaching obligations
- PhD co-supervision
- dynamic interdisciplinary team in one of Germany’s top research universities

https://tinyurl.com/jhxb4wu2

Here is material for an intermediate course on #Bayesian #regression modeling using #brms :...

Here is material for an intermediate course on #Bayesian #regression modeling using #brms : https://michael-franke.github.io/Bayesian-Regression/

Covers:

- prior / posterior model checking
- generalized linear & non-linear models
- mixture and distributional models
- MCMC methods (some Stan)
- model comparison
- causal inference

My first serious steps with #quarto (#RStudio) for publishing: really, really nice!Pros: clean syntax, live rendering during dev (w/o...

My first serious steps with #quarto (#RStudio) for publishing: really, really nice!

Pros: clean syntax, live rendering during dev (w/o massive recompile of whole project), slick cmd line interface.

Con: emacs quarto mode has not yet bended to my will.

Computing Bayes Factors is difficult in general. But for issues like "parameter > number", there is a nice method:...

Computing Bayes Factors is difficult in general. But for issues like "parameter > number", there is a nice method: _encompassing models_. It's simple, elegant and powerful, i.o.w., it's beautifully Bayesian. 🤩

Here is an overview: tinyurl.com/3pxs2pc9

#BDA #Bayes #BayesianDataAnalysis #ml #stats

New paper on the production of verbal expressions of (uncertain) probability #xprag. Using #BayesianDataAnalysis to investigate structural...

New paper on the production of verbal expressions of (uncertain) probability #xprag. Using #BayesianDataAnalysis to investigate structural properties of the latent #semantics of expressions like "probable" or "low chance" in a probabilistic model of #pragmatic utterance choice.
Joint w/ Bob van Tiel & Uli Sauerland. #openaccess
https://doi.org/10.1162/opmi_a_00066

According to high-N observational evidence (my father's record collection), Ray Brown played bass on /every/ Jazz album from the 50s and...

According to high-N observational evidence (my father's record collection), Ray Brown played bass on /every/ Jazz album from the 50s and 60s.

Christina Pawlowitsch is organizing an exciting event on "Applications of Game Theory in the Study of...

Christina Pawlowitsch is organizing an exciting event on "Applications of Game Theory in the Study of Language".

https://game-theory.u-paris2.fr/

Great combination of speakers from different fields.

There is a call for contributions. -> Submit your own work by February 28th or just join the fun in Paris next summer.

"The way in which linguists show affection for a theory is by criticizing it."[wise person, off-hand, during a project meeting,...

"The way in which linguists show affection for a theory is by criticizing it."
[wise person, off-hand, during a project meeting, just now]

witness: @rdhawkins

#linguistics
#theory
folk #philosophyofscience

Unbelievable! Inbox zero day. First time in like 3 years.

Unbelievable! Inbox zero day. First time in like 3 years.

For some #ML papers with cool ideas and thousands of citations, if that was a student paper in my class, I would massively subtract points...

For some #ML papers with cool ideas and thousands of citations, if that was a student paper in my class, I would massively subtract points for shabby notation and imprecise terminology. And would not stop ranting about uninformative graphs, abundant typos, and dodgy performance comparisons. Is this a peacock's feather: I'm super-smart, so I thrive even if I cannot communicate?

What's the name of this illness I definitely own? - I cannot just use #msword. Completely impossible. Whenever #msword is required...

What's the name of this illness I definitely own? - I cannot just use #msword. Completely impossible. Whenever #msword is required (bloody hell), I HAVE TO first spent a ridiculous amount of time #procrastinating. Achieve something new, something non-standard. Like using vim-style navigation inside #msoffice or streamlining conversion org -> odt -> doc. I can simply NOT just use the software w/o anything that reassures me: "Don't worry, kid! You are still in control!" - Crazy!

Getting started here. Expect me to talk about #stats, #bayes, #brms, #mcmc_stan, #probprag, #emacs, #cogsci, #cogmod, #pragmatics and more.

Getting started here. Expect me to talk about #stats, #bayes, #brms, #mcmc_stan, #probprag, #emacs, #cogsci, #cogmod, #pragmatics and more.