micha heilbron
@mheilbron.bsky.social
Assistant Professor of Cognitive AI @UvA Amsterdam
language and vision in brains & machines
cognitive science 🤝 AI 🤝 cognitive neuroscience
michaheilbron.github.io
language and vision in brains & machines
cognitive science 🤝 AI 🤝 cognitive neuroscience
michaheilbron.github.io
This is, without a doubt, the best popular article about current state of AI. And on whether LLMs are truly 'thinking' or 'understanding' -- and what that question even means
www.newyorker.com/magazine/202...
www.newyorker.com/magazine/202...
The Case That A.I. Is Thinking
ChatGPT does not have an inner life. Yet it seems to know what it’s talking about.
www.newyorker.com
November 7, 2025 at 10:32 AM
This is, without a doubt, the best popular article about current state of AI. And on whether LLMs are truly 'thinking' or 'understanding' -- and what that question even means
www.newyorker.com/magazine/202...
www.newyorker.com/magazine/202...
New paper on memorability, with @davogelsang.bsky.social !
New preprint out together with @mheilbron.bsky.social
We find that a stimulus' representational magnitude—the L2 norm of its DNN representation—predicts intrinsic memorability not just for images, but for words too.
www.biorxiv.org/content/10.1...
We find that a stimulus' representational magnitude—the L2 norm of its DNN representation—predicts intrinsic memorability not just for images, but for words too.
www.biorxiv.org/content/10.1...
Representational magnitude as a geometric signature of image and word memorability
What makes some stimuli more memorable than others? While memory varies across individuals, research shows that some items are intrinsically more memorable, a property quantifiable as “memorability”. ...
www.biorxiv.org
September 18, 2025 at 10:45 AM
New paper on memorability, with @davogelsang.bsky.social !
Reposted by micha heilbron
New preprint out together with @mheilbron.bsky.social
We find that a stimulus' representational magnitude—the L2 norm of its DNN representation—predicts intrinsic memorability not just for images, but for words too.
www.biorxiv.org/content/10.1...
We find that a stimulus' representational magnitude—the L2 norm of its DNN representation—predicts intrinsic memorability not just for images, but for words too.
www.biorxiv.org/content/10.1...
Representational magnitude as a geometric signature of image and word memorability
What makes some stimuli more memorable than others? While memory varies across individuals, research shows that some items are intrinsically more memorable, a property quantifiable as “memorability”. ...
www.biorxiv.org
September 18, 2025 at 9:53 AM
New preprint out together with @mheilbron.bsky.social
We find that a stimulus' representational magnitude—the L2 norm of its DNN representation—predicts intrinsic memorability not just for images, but for words too.
www.biorxiv.org/content/10.1...
We find that a stimulus' representational magnitude—the L2 norm of its DNN representation—predicts intrinsic memorability not just for images, but for words too.
www.biorxiv.org/content/10.1...
New preprint! w/@drhanjones.bsky.social
Adding human-like memory limitations to transformers improves language learning, but impairs reading time prediction
This supports ideas from cognitive science but complicates the link between architecture and behavioural prediction
arxiv.org/abs/2508.05803
Adding human-like memory limitations to transformers improves language learning, but impairs reading time prediction
This supports ideas from cognitive science but complicates the link between architecture and behavioural prediction
arxiv.org/abs/2508.05803
Human-like fleeting memory improves language learning but impairs reading time prediction in transformer language models
Human memory is fleeting. As words are processed, the exact wordforms that make up incoming sentences are rapidly lost. Cognitive scientists have long believed that this limitation of memory may, para...
arxiv.org
August 18, 2025 at 12:40 PM
New preprint! w/@drhanjones.bsky.social
Adding human-like memory limitations to transformers improves language learning, but impairs reading time prediction
This supports ideas from cognitive science but complicates the link between architecture and behavioural prediction
arxiv.org/abs/2508.05803
Adding human-like memory limitations to transformers improves language learning, but impairs reading time prediction
This supports ideas from cognitive science but complicates the link between architecture and behavioural prediction
arxiv.org/abs/2508.05803
CCN has arrived here here in Amsterdam!
Come find me to meet or catch up
Some highlights from students and collaborators:
Come find me to meet or catch up
Some highlights from students and collaborators:
August 12, 2025 at 11:14 AM
CCN has arrived here here in Amsterdam!
Come find me to meet or catch up
Some highlights from students and collaborators:
Come find me to meet or catch up
Some highlights from students and collaborators:
Waarom vergeet je namen maar weet je nog precies wat iemand doet? En zijn herinneringen ooit echt helemaal weg?
Ik ging bij Oplossing Gezocht in gesprek over hoe ons brein informatie opslaat en waarom vergeten eigenlijk heel slim is:
www.nemokennislink.nl/publicaties/...
Ik ging bij Oplossing Gezocht in gesprek over hoe ons brein informatie opslaat en waarom vergeten eigenlijk heel slim is:
www.nemokennislink.nl/publicaties/...
Waarom kom ik toch niet op die naam?
Eh, je weet wel wie... dinges! Heb jij ook soms zo’n moeite om op een naam te komen? Hersenonderzoeker Micha Heilbron legt uit hoe dat komt - en waarom een naam eigenlijk niet zo belangrijk is.
www.nemokennislink.nl
July 15, 2025 at 8:47 AM
Waarom vergeet je namen maar weet je nog precies wat iemand doet? En zijn herinneringen ooit echt helemaal weg?
Ik ging bij Oplossing Gezocht in gesprek over hoe ons brein informatie opslaat en waarom vergeten eigenlijk heel slim is:
www.nemokennislink.nl/publicaties/...
Ik ging bij Oplossing Gezocht in gesprek over hoe ons brein informatie opslaat en waarom vergeten eigenlijk heel slim is:
www.nemokennislink.nl/publicaties/...
Reposted by micha heilbron
Exciting new preprint from the lab: “Adopting a human developmental visual diet yields robust, shape-based AI vision”. A most wonderful case where brain inspiration massively improved AI solutions.
Work with @zejinlu.bsky.social @sushrutthorat.bsky.social and Radek Cichy
arxiv.org/abs/2507.03168
Work with @zejinlu.bsky.social @sushrutthorat.bsky.social and Radek Cichy
arxiv.org/abs/2507.03168
arxiv.org
July 8, 2025 at 1:04 PM
Exciting new preprint from the lab: “Adopting a human developmental visual diet yields robust, shape-based AI vision”. A most wonderful case where brain inspiration massively improved AI solutions.
Work with @zejinlu.bsky.social @sushrutthorat.bsky.social and Radek Cichy
arxiv.org/abs/2507.03168
Work with @zejinlu.bsky.social @sushrutthorat.bsky.social and Radek Cichy
arxiv.org/abs/2507.03168
Reposted by micha heilbron
New preprint, w/ @predictivebrain.bsky.social !
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
Higher-level spatial prediction in natural vision across mouse visual cortex
Theories of predictive processing propose that sensory systems constantly predict incoming signals, based on spatial and temporal context. However, evidence for prediction in sensory cortex largely co...
www.biorxiv.org
May 23, 2025 at 11:39 AM
New preprint, w/ @predictivebrain.bsky.social !
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
New preprint, w/ @predictivebrain.bsky.social !
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
Higher-level spatial prediction in natural vision across mouse visual cortex
Theories of predictive processing propose that sensory systems constantly predict incoming signals, based on spatial and temporal context. However, evidence for prediction in sensory cortex largely co...
www.biorxiv.org
May 23, 2025 at 11:39 AM
New preprint, w/ @predictivebrain.bsky.social !
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
we've found that visual cortex, even when just viewing natural scenes, predicts *higher-level* visual features
The aligns with developments in ML, but challenges some assumptions about early sensory cortex
www.biorxiv.org/content/10.1...
Reposted by micha heilbron
Higher-level spatial prediction in natural vision across mouse visual cortex https://www.biorxiv.org/content/10.1101/2025.05.15.654212v1
May 20, 2025 at 1:15 AM
Higher-level spatial prediction in natural vision across mouse visual cortex https://www.biorxiv.org/content/10.1101/2025.05.15.654212v1
Reposted by micha heilbron
"Opoffering Psychologie door Universiteiten van Nederland (UNL) raakt ons allemaal. (...) Dit is geen compromis, maar een ‘verdeel-en-heers’-strategie. Vandaag is het Psychologie, morgen een andere opleiding." #DOEHETNIET @woinactie.bsky.social www.scienceguide.nl/2025/04/coll...
‘Collega’s, opoffering Psychologie door UNL raakt ons allemaal’ - ScienceGuide
De respons van UNL op de WIB is willekeurig en ondemocratisch, betogen vier Amsterdamse onderzoekers.
www.scienceguide.nl
April 23, 2025 at 12:23 PM
"Opoffering Psychologie door Universiteiten van Nederland (UNL) raakt ons allemaal. (...) Dit is geen compromis, maar een ‘verdeel-en-heers’-strategie. Vandaag is het Psychologie, morgen een andere opleiding." #DOEHETNIET @woinactie.bsky.social www.scienceguide.nl/2025/04/coll...
Collega's in andere disciplines: spreek je uit tegen de UNL opoffering van psychologie-opleidingen
Vandaag wij, morgen jullie? Laat je stem horen bij je bestuur.
Met @mkdeserno.bsky.social @bnbakker.bsky.social @berendvdkolk.bsky.social @scienceguide.bsky.social
www.scienceguide.nl/2025/04/coll...
Vandaag wij, morgen jullie? Laat je stem horen bij je bestuur.
Met @mkdeserno.bsky.social @bnbakker.bsky.social @berendvdkolk.bsky.social @scienceguide.bsky.social
www.scienceguide.nl/2025/04/coll...
‘Collega’s, opoffering Psychologie door UNL raakt ons allemaal’ - ScienceGuide
De respons van UNL op de WIB is willekeurig en ondemocratisch, betogen vier Amsterdamse onderzoekers.
www.scienceguide.nl
April 23, 2025 at 11:25 AM
Collega's in andere disciplines: spreek je uit tegen de UNL opoffering van psychologie-opleidingen
Vandaag wij, morgen jullie? Laat je stem horen bij je bestuur.
Met @mkdeserno.bsky.social @bnbakker.bsky.social @berendvdkolk.bsky.social @scienceguide.bsky.social
www.scienceguide.nl/2025/04/coll...
Vandaag wij, morgen jullie? Laat je stem horen bij je bestuur.
Met @mkdeserno.bsky.social @bnbakker.bsky.social @berendvdkolk.bsky.social @scienceguide.bsky.social
www.scienceguide.nl/2025/04/coll...
Reposted by micha heilbron
✨New paper ✨
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
April 7, 2025 at 2:56 PM
✨New paper ✨
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
en route to Scotland, visiting universities of Edinburgh and Glasgow tomorrow & Friday.
do reach out if you are around and want to talk language models, brains, or anything in between!
do reach out if you are around and want to talk language models, brains, or anything in between!
March 19, 2025 at 5:15 PM
en route to Scotland, visiting universities of Edinburgh and Glasgow tomorrow & Friday.
do reach out if you are around and want to talk language models, brains, or anything in between!
do reach out if you are around and want to talk language models, brains, or anything in between!
lovely and slightly surreal to be back at Queen Square after all those years, but had a blast talking about generative AI and the predictive brain at the FIL today @peterkok.bsky.social @clarepress.bsky.social
March 19, 2025 at 3:12 PM
lovely and slightly surreal to be back at Queen Square after all those years, but had a blast talking about generative AI and the predictive brain at the FIL today @peterkok.bsky.social @clarepress.bsky.social
Reposted by micha heilbron
Encouraging to see 5000+ protesters at our local Amsterdam protests against the budget cuts in higher education in the NL! In this time of geopolitical tensions, changing climate, democratic backsliding, & tech revolutions, governments should *invest* in knowledge & the future of the next generation
March 17, 2025 at 2:36 PM
Encouraging to see 5000+ protesters at our local Amsterdam protests against the budget cuts in higher education in the NL! In this time of geopolitical tensions, changing climate, democratic backsliding, & tech revolutions, governments should *invest* in knowledge & the future of the next generation
Reposted by micha heilbron
We are excited that Ellie Pavlick from Brown University & Google DeepMind will be the third keynote speaker of CCN2025!
Join us in August for the 8th CCN conference, on cognitive science, neuroscience, and AI. Details on paper submission (both short & full length papers) below!
Join us in August for the 8th CCN conference, on cognitive science, neuroscience, and AI. Details on paper submission (both short & full length papers) below!
January 24, 2025 at 1:00 PM
We are excited that Ellie Pavlick from Brown University & Google DeepMind will be the third keynote speaker of CCN2025!
Join us in August for the 8th CCN conference, on cognitive science, neuroscience, and AI. Details on paper submission (both short & full length papers) below!
Join us in August for the 8th CCN conference, on cognitive science, neuroscience, and AI. Details on paper submission (both short & full length papers) below!
Cog-sci folks!
What's your favourite platform for developing/running online studies? I’m looking for something that: 1) integrates with prolific; 2) requires minimal web-development experience, ideally just python.
January 21, 2025 at 12:20 PM
Cog-sci folks!
What's your favourite platform for developing/running online studies? I’m looking for something that: 1) integrates with prolific; 2) requires minimal web-development experience, ideally just python.
Reposted by micha heilbron
CCN 2025 Area Chair Nomination
Intrigued about supporting CCN’s new 8-page conference proceedings track? 📄
Fill out this form to nominate yourself or a colleague to act as (Senior) Area Chair at CCN25.
forms.gle/MTBgZZxQaDY8...
Intrigued about supporting CCN’s new 8-page conference proceedings track? 📄
Fill out this form to nominate yourself or a colleague to act as (Senior) Area Chair at CCN25.
forms.gle/MTBgZZxQaDY8...
CCN 2025 Area Chair Nomination
Area Chairs nominated to serve for the 8-page conference proceedings track at CCN 2025 should meet the following requirements:
be a senior researcher (professor, research fellow, postdoc);
have author...
forms.gle
January 10, 2025 at 9:55 AM
CCN 2025 Area Chair Nomination
Intrigued about supporting CCN’s new 8-page conference proceedings track? 📄
Fill out this form to nominate yourself or a colleague to act as (Senior) Area Chair at CCN25.
forms.gle/MTBgZZxQaDY8...
Intrigued about supporting CCN’s new 8-page conference proceedings track? 📄
Fill out this form to nominate yourself or a colleague to act as (Senior) Area Chair at CCN25.
forms.gle/MTBgZZxQaDY8...
Reposted by micha heilbron
Key-value memory is an important concept in modern machine learning (e.g., transformers). Ila Fiete, Kazuki Irie, and I have written a paper showing how key-value memory provides a way of thinking about memory organization in the brain:
arxiv.org/abs/2501.02950
arxiv.org/abs/2501.02950
Key-value memory in the brain
Classical models of memory in psychology and neuroscience rely on similarity-based retrieval of stored patterns, where similarity is a function of retrieval cues and the stored patterns. While parsimo...
arxiv.org
January 7, 2025 at 9:21 AM
Key-value memory is an important concept in modern machine learning (e.g., transformers). Ila Fiete, Kazuki Irie, and I have written a paper showing how key-value memory provides a way of thinking about memory organization in the brain:
arxiv.org/abs/2501.02950
arxiv.org/abs/2501.02950
Reposted by micha heilbron
Felix Hill was such an incredible mentor — and occasional cold water swimming partner — to me. He's a huge part of why I joined DeepMind and how I've come to approach research. Even a month later, it's still hard to believe he's gone.
January 2, 2025 at 7:01 PM
Felix Hill was such an incredible mentor — and occasional cold water swimming partner — to me. He's a huge part of why I joined DeepMind and how I've come to approach research. Even a month later, it's still hard to believe he's gone.
Reposted by micha heilbron
Why do diverse ANNs resemble brain representations? Check out our new paper with Colton Casto, @nogazs.bsky.social , Colin Conwell, Mark Richardson, & @evfedorenko.bsky.social on “Universality of representation in biological and artificial neural networks.” 🧠🤖
tinyurl.com/yckndmjt
tinyurl.com/yckndmjt
Universality of representation in biological and artificial neural networks
Many artificial neural networks (ANNs) trained with ecologically plausible objectives on naturalistic data align with behavior and neural representations in biological systems. Here, we show that this...
tinyurl.com
December 27, 2024 at 8:14 PM
Why do diverse ANNs resemble brain representations? Check out our new paper with Colton Casto, @nogazs.bsky.social , Colin Conwell, Mark Richardson, & @evfedorenko.bsky.social on “Universality of representation in biological and artificial neural networks.” 🧠🤖
tinyurl.com/yckndmjt
tinyurl.com/yckndmjt
‘people who are paranoid or perceive meaning in the universe’ is my new favourite demographic
Out today: we discovered a new class of social hallucination - perception of chasing with high confidence. It is elevated in people who are paranoid and people who perceive meaning in the universe www.nature.com/articles/s44...
Paranoid and teleological thinking give rise to distinct social hallucinations in vision - Communications Psychology
When asked to judge if a chase was present in a visual display of moving discs, people with higher paranoia and teleological thinking were more likely to perceive a chase in its absence. They were als...
www.nature.com
December 18, 2024 at 11:54 AM
‘people who are paranoid or perceive meaning in the universe’ is my new favourite demographic
Reposted by micha heilbron
Column | De wooncrisis is misschien wel de belangrijkste oorzaak van het chagrijn in Nederland, maar er gebeurt niks of zelfs precies het verkeerde.
Bouwen moet, maar om wonen weer betaalbaar te maken zijn andere maatregelen nodig
De wooncrisis is misschien wel de belangrijkste oorzaak van het chagrijn in Nederland, maar er gebeurt niks of zelfs precies het verkeerde.
buff.ly
December 17, 2024 at 7:34 AM
Column | De wooncrisis is misschien wel de belangrijkste oorzaak van het chagrijn in Nederland, maar er gebeurt niks of zelfs precies het verkeerde.