Carl Hodgetts
@cjhodgetts.bsky.social
Cognitive neuroscientist at @RHULPsychology. Interested in how we represent and navigate the world around us. Hippocampus fanboy 🍤 #firstgen
This is wild - great work!
June 12, 2025 at 3:07 PM
This is wild - great work!
We hope that this work proves useful for researchers in identifying where and when these limitations are most consequential.
@rhulpsychology.bsky.social @birkbeckpsychology.bsky.social
@rhulpsychology.bsky.social @birkbeckpsychology.bsky.social
May 16, 2025 at 11:40 AM
We hope that this work proves useful for researchers in identifying where and when these limitations are most consequential.
@rhulpsychology.bsky.social @birkbeckpsychology.bsky.social
@rhulpsychology.bsky.social @birkbeckpsychology.bsky.social
In this paper/commentary, Ulrike Hahn and I show that comparisons between even very simple geometric stimuli (like 🟧🔵!) can expose serious issues with these models - for example, by assuming similarities that do not fit neatly into Euclidean space.
May 16, 2025 at 11:40 AM
In this paper/commentary, Ulrike Hahn and I show that comparisons between even very simple geometric stimuli (like 🟧🔵!) can expose serious issues with these models - for example, by assuming similarities that do not fit neatly into Euclidean space.
The idea that we represent concepts in multi-dimensional “cognitive spaces” has been hugely influential (and is having a bit of comeback in psych/neuro). But these spaces have built-in topological constraints that limit their utility - both practically and theoretically.
May 16, 2025 at 11:40 AM
The idea that we represent concepts in multi-dimensional “cognitive spaces” has been hugely influential (and is having a bit of comeback in psych/neuro). But these spaces have built-in topological constraints that limit their utility - both practically and theoretically.
Sounds fantastic - looking forward to hearing more!
February 7, 2025 at 4:00 PM
Sounds fantastic - looking forward to hearing more!