It's time that linguistics seriously discusses formal training. We are a field that widely uses formal/mathematical tools (formal language theory, logic, probability theory, statistics), yet we do not provide strong systematic training of these tools to our students. why is that?
Exploring a new "Template Expression" API design for symbolic regression that allows joint optimization of symbolic expressions with parameter vectors. Can be used for optimizing a neural network simultaneously with a symbolic model, finding parametric expressions, learning constants in fixed forms, and much more.
Would love feedback on the design!
github.com/MilesCranmer/Symb…
We aren't the only primates who appreciate a good joke. New piece on the origins of humor by DISI Director Erica Cartmill!
scientificamerican.com/artic…
My paper on hierarchical plans is out in Cognition!🎉
tldr: We ask participants to generate hierarchical plans in a programming game. People prefer to reuse beyond what standard accounts predict, which we formalize as induction of a grammar over actions.
authors.elsevier.com/a/1kBQr…
👩🎓👨🎓Fully funded PhD position available to come work with me at @UniOslo using iterative learning experiments to understand the evolution of sound symbolism.
🔴Deadline is 12. Jan '24
⏲️Desired starting date is Mar/April '24
jobbnorge.no/en/available-jo…
(1/5) Very excited to announce the publication of Bayesian Models of Cognition: Reverse Engineering the Mind. More than a decade in the making, it's a big (600 pages) beautiful book covering both the basics and recent work: mitpress.mit.edu/97802620494…
"we don't know what to do with those chains yet, but I have the gut feeling that once we do, probabilistic programming will have another breakthrough" - @remilouf on parallelizing 100s of MCMC chains with JAX
Similarly, there are several research directions today seen as long-abandoned failures, that are only waiting for the right amount of attention and the right level of compute scale to shine. Genetic algorithms are one of them.
New pre-print (tx @ChrisMMCox!): The social context of turn-taking osf.io/preprints/psyarxiv/zw… Comments welcome!
TL;DR: Response latency in conversation depends on a multidimensional interplay of individual differences, turn-by turn information flow & contextual affordances.
'A Drive to Survive' is available for pre-order!
mitpress.mit.edu/97802625513…
And while it doesn't yet have a cover it does have some embarrassingly nice blurbs from @evantthompson and @CogsAndy, two of the philosophers who most inspired its content.
Preprint from myself and Daphne Demekas. We propose a mechanism for synthetic data generation within active inference and present simulations to explore its implications for understanding states of 'pure consciousness'
osf.io/preprints/psyarxiv/w6…
The traditional theoretical question of whether maths is discovered or invented has a counterpart in the ambiguity of the *experience* of reasoning, which can feel like perception (of an already present structure) or action (building up the structure).
(Who knows what Derrida really wants to say, but) to me this sounds like the former - though he calls it an "apparatus, or machine" rather than structure: youtu.be/BSsDRf2wnOk?si=AEO9…
And he describes this sense of it happening *to* him. Sometimes reasoning feels more like following a pattern with your eyes, with a large component of attentional direction, and sometimes more like smelling or hearing, happening spontaneously and even intrusively.
@UniOslo_RITMO is recruiting a PhD student to study musical microrhythm, supervised by @annedanielsen and Rainer Polack. This is the best imaginable situation for a budding rhythm nerd. If you have any star undergrads in your lab open to a move to Norway...jobbnorge.no/en/available-jo…