It is an amazing time to work in the cognitive science of language. Here are a few remarkable recent results, many of which highlight ways in which the critiques of LLMs (especially from generative linguistics!) have totally fallen to pieces. One claim was that LLMs can't be right because they learn "impossible languages." This was never really justified, and now @JulieKallini and collaborators show its probably not true:

https://x.com/JulieKallini/status/1746992945738526985?s=20

I love it already.

https://twitter.com/GlassHealthHQ/status/1620092094034620421

Really great job here, @GlassHealthHQ. Your technology is ready for clinical use.

Yes, ChatGPT is amazing and impressive. No, @OpenAI has not come close to addressing the problem of bias. Filters appear to be bypassed with simple tricks, and superficially masked.

And what is lurking inside is egregious.

@Abebab @sama
tw racism, sexism. It's not a fluke

still just unbelievable

Yeah, yeah, quantum mechanics and relativity are counterintuitive because we didn’t evolve to deal with stuff on those scales.

But more ordinary things like numbers, geometry, and procedures are also baffling. Here’s a little 🧵 on weird truths in math. My favorite example – the Banach-Tarski paradox – shows how you can cut a sphere into a few pieces (well, sets) and then re-assemble the pieces into TWO IDENTICAL copies of the sphere you started with.

Everyone seems to think it's absurd that large language models (or something similar) could show anything like human intelligence and meaning. But it doesn’t seem so crazy to me. Here's a dissenting 🧵 from cognitive science. The news, to start, is that this week software engineer @cajundiscordian was placed on leave for violating Google's confidentiality policies, after publicly claiming that a language model was "sentient"
nytimes.com/2022/06/12/tec…

I am sooooooo excited for this paper. We've spent years developing a super fast program induction library. We use it to learn key pieces of language structure.

So much of what Chomskyan linguists say about learnability is totally wrong.

🧵

pnas.org/content/119/5/… We show that a program-learning model can construct grammars and other computational devices from just observing utterances (sentences). It takes just a tiny amount of data to learn key patterns and structures from natural language which have been argued to be innate/unlearnable

Did you know that there are psychologists who study "stereotype accuracy"? I've always wondered what the hell, so I've been reading it recently. For the record, it’s exactly as bad as it sounds.

Here’s a thread. I started with @PsychRabble et al's review paper, "Stereotype accuracy: one of the largest and most replicable effects in all of social psychology." It claims the correlation between stereotypes and reality is stronger than most effects in psychology.
gwern.net/docs/psycholog…

The GameStop $GME news is too funny not to write a thread about.

It's also the most interesting intersection of psychology and politics going on. I think the best overview of what's happening is this NYT article:
nytimes.com/2021/01/27/bus…

A new paper on (i) how to connect high-level cognitive theories to neuroscience / neural nets, and (ii) how learners can construct new concepts like number or logic without presupposing them. "Church encoding" is the metaphor we need.
link.springer.com/article/10.100… Kids learn concepts like number or logic, but it's not really clear how that's actually possible: what could it mean for a learner (or computer!) to *not* have numbers or logic? Can you build a computer without them? And what do these abstract things mean on a neural level?

This article by @JeffFynnPaul has been going around. He argues that it’s a “myth” that Europeans took land from Native Americans. @cakrolik and I read it and it’s one of the worst argued pieces we’ve ever seen. Here is our thread.
spectator.co.uk/article/the-my… The overall claim of his article is that liberals have perpetuated a “Myth of the Stolen Country” that the US “was founded by a monumental act of genocide, accompanied by larceny on the grandest scale.”

I am so excited that this new paper with @samisaguy is out. We explain how humans perceive number by mathematically deriving an optimal perceptual system with bounded information processing. Four N=100 pre-registered experiments match the predictions.
nature.com/articles/s4156… People have long hypothesized that there are two systems of number perception: an exact system that lets us perfectly perceive numbers less than about 4, and an approximate one for large numbers.

Super excited to talk tomorrow (July 30th, 3pm pacific) at Abralin ao Vivo about joint with Yuan Yang. I'll be presenting a long-running project on language acquisition that tackles language learnability questions with Bayesian program learning tools.

Here's a summary thread.

https://twitter.com/abralin_oficial/status/1288465678215897090

This is part of an amazing remote talk series by @abralin_oficial presenting language work all summer
abralin.org/site/en/evento…

New result published by @SpringerNature has proven mathematically that homeopathy works.

I had to tweet about this paper. Homeopathy is often thought to be "natural medicine." In truth, it's not anything medicine. Homeopathy is actually a remarkable delusion: the idea is that you take a substance and dilute it in water until there are no molecules of the substance left.
en.wikipedia.org/wiki/Homeopathy

The ideas in the @Harpers letter were destroyed on Twitter yesterday. Here's a thread/meta-thread. Yesterday, a variety of public figures, from @jk_rowling to Malcolm @Gladwell to @SalmanRushdie signed a letter in Harper’s saying that "The free exchange of information and ideas, the lifeblood of a liberal society, is daily becoming more constricted."
harpers.org/a-letter-on-ju…

Here is why IQ is bullshit.

A thread. A recent paper in @PsychScience relied on national differences in IQ that held that the average IQ of some countries is below the threshold for intellectual disability.

"In publishing an Op-Ed that appears to call for violence, promotes hate, and rests its arguments on several factual inaccuracies while glossing over other matters that require—and were not met with—expert legal interpretation, we fail our readers"
slate.com/news-and-polit… I still can't get over how bad the @nytimes editorial page editor @JBennet is at making an argument.

https://twitter.com/JBennet/status/1268328278730866689?s=20

So @BabyMiroTweets, @celestekidd, and I got a little microscope to look at stuff while sheltering in place. We learned not to go outside because THERE ARE THESE SWIMMING IN THE WATER RIGHT OUTSIDE OUR HOUSE

My and @celestekidd's labs @BerkeleyPsych are seeking a postdoctoral researcher to lead a large developmental study on the relationship between early mathematics and attention. Candidates must have strong quantitative skills (e.g. programming, statistics, experimental design) and preferably will have background in developmental studies and/or eye-tracking experiments.

I would say this story is unbelievable but it's not -- the history of who invented using channelrhodopsin in neuroscience. statnews.com/2016/09/01/opt… "... Pan wrote to the editor at @NatureNeuro asking how they could have rejected his paper but published Boyden’s. In her response, the editor replied that while the papers were similar ... Pan’s paper was too narrow, only focusing on using channelrhodopsin to restore vision..."

At the end of her #NeurIPS2019 talk on how we come to know about the world, @celestekidd takes a moment to comment on the "climate for men" that #MeToo has created and what people don't know about it. @NeurIPSConf @WomenInCogSci #WiML2019 @AnnOlivarius @ucberkely @BerkeleyPsych In science and tech, "people believe that men are being fired for subtle comments or minor jokes, or just plain kindness or cordial behavior. This perception makes people very nervous. What I want to say today to all of the men in the room is that you have been misled."