Judges' asylum decisions vary wildly, violating a "prerequisite for ...legal coherence—to 'treat like cases alike."
This paper shows how #AI and #algorithms can "instantiate a reasonable approximation of a coherence theory of #truth":
Recognizing the fallibilities of #ChatGPT was most reliably predicted by
(a) better performance on a reflection test
(b) more self-reported prioritization of arguments and evidence (over myside political #bias)
in a study of 999 students: https://doi.org/10.1080/10494820.2025.2454448
The Hitchhiker's Guide to Digital Colour https://lobste.rs/s/ibfmsg #art #cogsci #graphics
https://hg2dc.com/
Thirty years ago, I published my PhD dissertation on Letter Spirit --- all about creativity, AI, and cognitive science. My advisor was Doug Hofstadter. Last week I was asked to give a keynote talk about that work in France. The talk was captured in video.
#AI #cogsci #creativity #typography
https://berryvilleiml.com/2025/02/26/revisiting-letter-spirit-thirty-years-later/
New on 'Stuff': Book Review - Taming Silicon Valley #cs
#cogsci
#neural-networks
#transformers
#llms
#llm
#book-review
#english
#️
️
️
️
️ https://stuff.graves.cl/posts/2025-03-01_17_03-book-review-taming-silicon-valley.html
Do people vacillate more before or after they can decide?
In this paper, most vacillations occurred before people could decide about logic or moral problems, except maybe for moral dilemmas.
Check out the #processTracing method: https://doi.org/10.1017/jdm.2024.15
The talk that I gave about Letter Spirit (my 30 year old thesis project) in France last week has been posted as a video. #AI #ML #cogsci #hofstadter #creativity #design #typography
ML/AI, Typographic Design, and the Four “I”s
During this talk I will touch on Intuition, Insight, and Inspiration. First I will set the context by introducing the Letter Spirit project and its microdomain — work I published exactly 30 years ago as a Ph.D. student of Doug Hofstadter’s...
https://berryvilleiml.com/2025/02/26/revisiting-letter-spirit-thirty-years-later/
The talk that I gave about Letter Spirit (my 30 year old thesis project) in France last week has been posted as a video. #AI #ML #cogsci #hofstadter #creativity #design #typography
ML/AI, Typographic Design, and the Four “I”s
During this talk I will touch on Intuition, Insight, and Inspiration. First I will set the context by introducing the Letter Spirit project and its microdomain — work I published exactly 30 years ago as a Ph.D. student of Doug Hofstadter’s. I will spend some time discussing the role of roles (and other mental structures) in creativity and human perception. Then I’ll take a quick run through the current state of “AI” (really ML) so we get a feeling of how LLMs actually work. We will talk about WHAT MACHINES and relate them to human cognition. Finally, just as I get around to intuition, insight, and inspiration, I will run out of time.
WOOP WOOP! Incoming ENSO Seminar!
Maria Clara Garavito Gomez will be presenting next week on "Intercorporeality at the limits of one's own body: An approach from somatoparaphrenia"
Thurs 6th March @ 1400 UTC
Check ensoseminars.com for time in your own timezone.
Abstract and event link here:
Can people's desire for control undermine the greater good?
Limiting the control people had in a dilemma's outcome seemed to make more of them choose to optimize their control, even though that resulted in saving fewer lives!
Revisiting Letter Spirit Thirty Years Later (my thesis with with Doug Hofstadter)
#AI #ML #cogsci #creativity #art #design #typography
https://berryvilleiml.com/2025/02/26/revisiting-letter-spirit-thirty-years-later/
Does #reading to #kids matter?
Reading aloud just one hour/day seemed to improve verbal and cognitive #intelligence in elementary and middle #school kids in two out of three cluster-randomized control trials in #Italy:
#STEM students outperformed #socialScience students not only on a #numeracy test, but also a mathematical #reflection test (N = 245).
Curiously, students' #ratioBias was not predicted by #reflection after controlling for #numeracy.
Language models trained to (a) generate longer solution steps and (b) expand their "inner monologue" to include reflective "self-corrections" performed 6-11% better on medical benchmarks.
Such #LLM #diagnosis seemed to employ a hypothetico-deductive method.
#introduction Hi, I'm Goji! I'm brand new to the #fediverse !
I'm currently pursuing a B.S. in #cogsci
Lately I've been spending my free time less on writing (I do short stories, and I'd like to self-publish a full novel) and more on reading.
I play some video games, but being a linux-boi has pared that back quite a bit.
My other hobbies are mountain biking, music, and succumbing to depression under a blanket.
You can expect me to talk about: #PoliticalReform, #AI, #foss, #books, university, #veganism, my perfect cats, and other interesting things I am learning about.
I will try to balance my emo, millennial nature with wit and well-written posts about things that matter. If I sound like a whiny bitch, stick around for a bit and see if it changes?
TWO new preprints from our lab on hierarchical processing in ferret auditory cortex!
Stay tuned for highlights on each study in the following posts
#Neuroscience #CogSci #BioRxiv