Cognitive scientist, psychologist, philosopher, author of Scientist in the Crib, Philosophical Baby, The Gardener & The Carpenter, WSJ Mind And Matter columnist
So delighted and honored to receive the Rumelhart prize for theoretical foundations of cognitive science, especially grateful because this is what I think my work has been about- bringing together philosophy, psychology and computation to figure out how knowledge is possible.
Huge congratulations to the brilliant ✨
@AlisonGopnik
,✨ recipient of the 2024 David E. Rumelhart Prize! 🏆🌟
Visit the
#RumelhartPrize
page to learn more
My new article with Emily Liquin out in Cognition, one of my favorite recent papers, 4 preregistered studies showing empirically that children explore, and learn, while adults exploit. So kids escape from avoidance "learning traps" that adults fall into.
@yudapearl
@GaryMarcus
@eliasbareinboim
@fivetroop
@ylecun
@vkhosla
Current LLM's with statistical next word prediction can't learn causal structure from experience like kids (see below). You might hack particular examples with prompt engineering and RLHF. But what the kids do requires radically different AI methods.
My latest in the WSJ based on Allyson Mackey and colleagues beautiful Nature Neuroscience Review paper. Good science that shows why programs supporting early childhood are so crucial.
Our latest paper, in Perspectives on Psychological Science, with Eunice Yiu and Eliza Kosoy, articulating the idea of Large AI Models as cultural technologies at more length and comparing and contrasting with human children
My latest in the Wall Street Journal: Reassuring new study shows that babies don't mind at all when Mom puts on a mask. What matters is the attunement and interaction between the baby and the person they're talking to, aka love.
New research led by
@LillianYuyan
reveals new ways in which early experiences influence how humans learn from and adapt to their changing circumstances
1/3
My paper in the new Phil Trans Life History and Learning Special issue, "Childhood as a solution to explore-exploit tensions" combining dev psych, computer science, neuroscience and evolution to explain why kids are so weird.
Beautiful paper in Nature H B from Giron and Wu lab supporting my "childhood as simulated annealing" idea, with great new findings, more generalization and directed exploration as well as randomness in kids, humans still outperform algorithms.
Exciting (free) CVPR AI Workshop: “Minds vs. Machines: How far are we from the common sense of a toddler?” June 15th, 2020. including me, J. Malik, L. Spelke, J. Tenenbaum, L. Smith, A. Oliva, B. Olshausen, L. Zitnick, and D. Yurovsky.
Are we basically selfish or generous ? Here's a link directly to the study that is the subject of my latest in the WSJ, fascinating research where people across multiple countries got 10,000.00 each. They spent much of it on others, including strangers.
My talk at the Simons Institute w new ideas on Large Language Models- they're cultural technologies like writing and print, accessing the knowledge of others rather than creating it. Plus expt showing that kids do better than AI in creating new knowledge
The actual copy of my new theory piece in Phil Trans -really pleased with this one which sums up new ideas integrating AI, evolutionary biology, and dev psych I've been working on for some time.
Really beautiful new study in Child Development from Mariel Goddu and Caren Walker showing that toddlers can simultaneously entertain multiple conflicting causal hypotheses. Simple , elegant method - profound results.
My latest theory/review "Childhood as a solution to explore-exploit tensions" drawing together evolutionary biology, AI, neuroscience and our dev psych studies to explain why kids, from crows to macaques to humans, are so weird-they explore, we exploit.
Study findings by Katherine Kimura &
@AlisonGopnik
indicate that young children can change their higher-order beliefs depending on the strength of the evidence presented to them and their prior beliefs:
#ChildDevelopment
Very nice new paper with Eric Brockbank, Caren Walker, Tania Lombrozo in Developmental Science "Ask me why, don't tell me why". Getting kids to explain leads to more abstract reasoning than pedagogical demonstration ... .
New paper with Willem Frankenhuis in Trends in Cognitive Sciences. Effects of early adversity on the development of explore-exploit trade-offs.
AddThis | Home
Many thanks for all the Guggenheim congrats! I'm using it to write a new book "Explore: A unified theory of childhood" aka a computational, cognitive and evolutionary explanation of why kids are so weird. And so pleased that Rebecca Saxe and Jennifer Doudna are fellow fellows..
I'm Alison Gopnik cognitive scientist and psychologist who studies how to make computers as good at learning as children. I also write the science column Mind and Matter for the Wall Street Journal.
#unapologeticallyfeministscientist
Very effective column on how to cut child poverty. Every developmental psychologist and neuroscientist will tell you just how transformative that would be
People often ask me for child development book recommendations. 📚
One of my very favorites, by Allison Gopnik, is called The Scientist in the Crib.
And the title couldn’t be more fitting.
While we sometimes associate the early years with “care” more than learning, this…
My latest in the WSJ, a striking study gave people 10,000.00 with no strings attached. Recipients gave much of it to others, even strangers, even when nobody knew what they'd done. Humans are as altruistic as they are selfish.
Netflix Babies series, new Season 2 Ep 6 on toddlers featuring our lab, my thoughts on theory of mind, and, best of all, cameo by beautiful grandchildren!
#netflixonbabies
,
Important new paper from Rebecca Zhu and me at Psych Science. 3yo can use metaphors to make new inferences, like scientists. Not quite learning that atomic structure is like the solar system but close!
What will happen when the Wizard waves her wand? New paper with Mariel Goddu, showing that preschoolers, contra conventional wisdom, understand very abstract relations, but only if Wizards make them happen!
A really excellent, balanced and thoughtful review of the debate over Large Language Models, understanding and intelligence from Melanie Mitchell
@melmitchell1
and David Krakauer
The two things I care about most are children and science, and both of them are in peril in this election. If you care too, vote for Biden/Harris and a Democratic senate now!
#VoteForScience
This is a terrific article, one of the best philosophical pieces on morality I've read - the rare case where a philosophical analysis is both insightful and sensible
Please join us in congratulating Kim Sterelny and Ben Fraser, whose work 'Evolution and Moral Realism' has been selected by the Philosopher's Annual as one of the ten best articles of 2017!
@TheBJPS
#fb
This paper, aka Wizards, is one of my favorite recent pubs. Psychologists have thought that 3 and 4 -yr olds couldn't use analogies or relations. But they can easily if they are the result of Wizardly transformations!
Nice piece in the NYT on the benefits of play, reporting on work by Daphna Buchsbaum, Caren Walker and me on pretend play and counterfactuals, as well as Stephanie Carlson (pretending to be Batman empowers you!) & Stahl and Feigenson on play in infants.
New paper with Azzura Ruggeri in Cognition - 3 and 4-year-olds selectively and rationally choosing the most informative experiments! Little scientists indeed.
New paper led by Azurra Ruggieri shows that toddlers spontaneously explore more as there is more relevant information to be gained. Getting into everything is experimentation! This is just what we'd like AIs to be able to do
This was one of my favorite book projects (especially the fictional sentimental correspondence between a snooty British formalist and salty NY empiricist in the intro). Full of profound ideas & increasingly relevant with the recent revival of causal learning in AI and ML
Delighted to hear from Pearl a Rumelhart (and Turing) winner. He's right, I wrote this chapter, to summarize work over 20 years, and be an approachable readable intro to the developmental/cogsci/computersci nexus for those who would like to know more.
This was probably the best interview I've ever done, so very glad to see it out there again, and the thread on the importance of paternal leave is great too.
This is something
@AlisonGopnik
says, and it's among the most important lines I've ever heard: We don't just care for others because we love them. We love them because we perform acts of care for them.
More on Large Language Models as "cultural technologies". They're not agents, intelligent or otherwise. Instead, LLM'S are like writing, print, libraries, and language itself, technologies that help humans learn from other humans.
I've admired Ezra Klein's writing about politics for a long time. But I was really impressed at his thoughtfulness, intelligence and range - this was one of the best radio/podcast/media conversations I've ever had, and just the kind of deep discussion about children we need.
I'm obsessed with
@AlisonGopnik
's idea that we don't provide care to people because we love them, we love them because of the acts of care we do for them. Parenthood makes this clearer, but looking back on my own life, it's true for so many relationships.
New essay on AI. GPT-3 etc are not individuals, smart or stupid, but like writing, print or internet search, they allow access to the knowledge of others (for good or ill) rather than creating new knowledge. Talk version here
Another great conversation with Ezra Klein, on childhood, exploration, evolution, AI, meditation, lantern consciousness, octopus selves, and - my favorite part- Mary Poppins as a dark Marvel Universe superhero
Here are three places where I talk about the LLM as cultural technology idea, WSJ is the most developed but behind a paywall. Arxiv paper is longest but still in submission.
Great talks in the Cogsci Society Machine and Human Common Sense Workshop now available on Youtube! Including mine on new work on causal functions and explore-exploit tradeoffs with M. Goddu and E. Liquin.
Just came across this lovely quote from Louise Gluck in her poem Nostos, which captures a lot of developmental psychology.
We look at the world once, in childhood.
The rest is memory.
So delighted to celebrate
@AlisonGopnik
who was awarded the APA Mentor Award yesterday. Such an incredibly deserving scientist and wonderful person.
A - We are so lucky to have you in our lives.
My latest in the WSJ, I really loved writing this, both as a memorial to the beloved Gretzky and as a chance to describe a really fascinating and clever study
Very cool new paper from our lab, in collaboration with
@uvhart
,
@ally_mackey
, accepted at Cognition. In a spontaneous creativity task kids explore more than adults and produce more unique outcomes.
My latest WSJ Column, proud to be an author on this new PNAS paper with Julia Kam, Zach Irving, Bob Knight and others - how the wandering mind reflects the wandering brain and why both are good for you!
And the original article
My latest in the WSJ on a beautiful elegantly designed new study by Munakata, Yanaoka et al in Psych Science. Lovely example of what we can learn from cross-cultural comparisons -- delay of gratification depends on cultural rituals and social rules
Videos of this fantastic cognitive science society workshop on explore/exploit trade-offs are now online. Remarkable convergences and interactions among developmental psychologists, evolutionary biologists, and...
If you're interested in more, this is a point made at length with lots more scientific examples, in my book The Gardener and the Carpenter: What the new science of child development tells us about the relations between parents and children.
Link to an important new article showing that policies like medicaid and child tax credits really can make a difference to young kids brain development and mental illness, and to my WSJ column about it.
New paper in Developmental Science w/ Ruggieri and E. Schulz preschoolers search longer when there is more information to be gained, inspiration for curiosity based AI systems.
SRCD is pleased to announce the recipients of the 2023 SRCD Biennial Awards! Join us in congratulating the following distinguished awardees at the SRCD Member Meeting and Awards Ceremony at
#SRCD23
! Read about the awardees:
More recent AI thinking: AI as cultural technology - like writing, print, libraries, internet search rather than individual intelligence. And similarities and differences to kids.
My latest column in the WSJ on the fascinating and clever new measure of creativity, the "creative foraging game" from
@uvhart
and colleagues at the Weizmann Institute
Preprint coming at PSA -new ideas about causal learning and intrinsic empowerment rewards, bridging Bayesian causal learning and reinforcement learning. TLDR hypothesis is that increasing empowerment necessarily increases causal learning and vice-versa.
"What is it like to be a baby?" Public talk by Alison Gopnik at the Melbourne State Library Sept 18
@MonashUni
@Monash_Arts
@cogphillab
(organised as part of Tim Bayne's exciting
@arc_gov_au
project on measuring consciousness). Come along!
Brief quote in the NYT this morning, for argument and evidence see new preprint. 1 LLM's are cultural technologies, like writing or print, not agents 2 We need careful dev psych methods to understand them, not anecdotes
Just finished season 2 of Lodge 49 which is 1) fantastically good 2) interesting, weird and sweet simultaneously 3) philosophically deep and also 4) little known even among my hippest friends and 5) so low in the ratings that it might not come back. A justification for tweeting!
The data babies use reflects their own active attempts to understand the world and their interactions with the people around them. The still very limited models may take advantage of this, but that doesn’t mean that they themselves have those abilities.
The
@KiddLab
and
@AlisonGopnik
labs at
@UCBerkeley
are seeking a postdoctoral researcher to lead research studies on the development of curiosity and learning throughout development, starting in early childhood.
So pleased by the strong response and interest in this piece. If you want to track the latest important research looking at AI and cognitive development follow coauthors Eliza Kosoy
@ElizaKosoy
and Eunice Yiu
@eunice_yiu_
.
School is important, but are students learning how to study and forgetting how to play?
“Rather than preschools being more like school, universities should be more like preschool,” says psych professor
@AlisonGopnik
. Full story:
@BerkeleyGse
#playmatters
Really proud of this one,the deep philosophical and psychological questions about care and the crucially urgent policy decisions.Shd be free with registration.
Lessons from lockdown: caregiving – Alison Gopnik on a revolution to properly value caregivers
My latest column in the WSJ on an important new study from
@rebecca_saxe
and
@danielnettle
in Cognition, left and right actually agree on redistribution principles. Article here
According to a new study, when it comes to our views on economic redistribution, being on the left or the right doesn’t matter as much as we think. via
@WSJ