Samuel Lippl retweeted
When do transformers length-generalize?
Generalizing to sequences longer than seen during training is a key challenge for transformers. Some tasks see success, others fail — but *why*? We introduce a theoretical framework to understand and predict length generalization.
Samuel Lippl retweeted
How does the brain bind action to value?
How does it navigate a tradeoff between this binding and generalization to novel situations?
Out today in Nature Neuro! rdcu.be/dUbMd
@JustFineNeuro @NeuroPolarbear @BecketEbs @neuromochi
See original 🧵 + updates below
How does the brain represent multiple different things at once in a single population of neurons? @JustFineNeuro, @NeuroPolarbear, @BecketEbs, @neuromochi and I show that it uses semi-orthogonal subspaces for each item.
Preprint here: arxiv.org/abs/2309.07766
Tweets below! (1/n)
Samuel Lippl retweeted
We live with the limitations of our memory, but don’t really know where they come from. Our new paper (doi.org/10.1073/pnas.2401032…) studies "swap errors", which we argue arise during memory manipulation – see thread for more! @timbuschman, @MattPanichello, @wjeffjohnston
Samuel Lippl retweeted
New preprint with @mBeiran on a theory of multiregion recurrent neural networks! We study networks where regions have a mix of random and structured connectivity, with low-rank “communication subspaces” between regions enabling selective activity routing.
arxiv.org/abs/2402.12188
Samuel Lippl retweeted
My work on dimensionality in random nonlinear RNNs, with Larry Abbott and Ashok Litwin-Kumar, is now out in @PhysRevLett!
Journal link: journals.aps.org/prl/abstrac…
PDF link: lk.zuckermaninstitute.columb…
Samuel Lippl retweeted
Ever wonder how the pianist can perform a new piece from sheet music, or effortlessly play a well-practiced sonata in a concert? We study the underlying neural circuits in ‘piano-playing’ rats and show that motor cortex and basal ganglia play different roles.
Samuel Lippl retweeted
Part 2 of our work (led by Kevin) on the role of motor cortex + basal ganglia in "flexible" vs. "automatic" tasks! Esp. excited about a phenomenon (which we model!) where subcortical consolidation of automatic behaviors is prevented by simultaneously practicing flexible behaviors
Samuel Lippl retweeted
Preprint alert🚨! In this new paper we study how humans decompose dynamical subprocesses and leverage the abstracted subprocesses for compositional reuse of experience in new situations.
psyarxiv.com/sxn4a/
Tweeprint to follow soon!
Samuel Lippl retweeted
NEW PAPER 🧵: Deep neural networks are complicated, but looking inside them ought to be simple. In this paper we introduce TorchLens, a package for extracting all activations and metadata from any #PyTorch model, and visualizing its structure, in as little as one line of code. 1/
Samuel Lippl retweeted
Excited to share triple (!) joint publications on bio-learning rules for the successor representation. With D. Aronov, L. Abbott, @e_mackevicius, we show RNNs can learn flexible predictive maps. Turns out two other groups were working on similar problems!
elifesciences.org/articles/8…
Samuel Lippl retweeted
Super excited to have this out! Hoping this will help pave the way for a theoretical understanding of the role of plastic synapses in neural computations. Full tweetprint forthcoming...
arxiv.org/abs/2302.08985
Samuel Lippl retweeted
I’m really excited to share that I will start my lab in 2023 at @epflSV in Lausanne as an ELISIR fellow! The lab will combine computational and experimental approaches to study the function of hippocampal-entorhinal circuits in memory formation! Interested in joining? ⬇️
Samuel Lippl retweeted
THAT'S DR GUCKY TO YOU!!! Congrats on defending @gucky92 !!!!!!
Samuel Lippl retweeted
You know that free-floating sense that multinational corporations are above the law, able to buy their way out of consequences for even the most blatant, heinous crimes?
There's a (nearly) unbelievable, highly concrete example of it underway right at this moment.
1/
Samuel Lippl retweeted
1/
There will be no update today.
At 8:30 am this morning, state police came into my house and took all my hardware and tech.
They were serving a warrant on my computer after DOH filed a complaint.
They pointed a gun in my face. They pointed guns at my kids..
Samuel Lippl retweeted
Is there anyone working on regulation protecting Ethical AI researchers, similar to whistleblower protection? Because with the amount of censorship & intimidation that goes on towards people in specific groups, how does anyone trust any real research in this area can take place?
Samuel Lippl retweeted
When Republicans talk about the Green New Deal being 100 trillion dollars, please know they‘re doing that Dr. Evil thing where they shout random, escalating numbers to sound ominous.
See for yourself. It’s written so you don’t need a law degree to read: congress.gov/116/bills/hres1…
Samuel Lippl retweeted
Moment to remember.
Q to Trump: Would you condemn white supremacists?
Trump doesn't answer the question & brings up ANTIFA.
Wallace: Will you condemn white supremacists?
Trump: "Proud Boys, stand back & stand by."
That is not a condemnation by any stretch of the imagination.
Samuel Lippl retweeted
ICE is now trying to deport students enrolled in colleges and universities that are teaching exclusively online due to COVID-19.
This is needlessly cruel and must be challenged in court.
This is bad. ICE just told students here on student visas that if their school is going online-only this fall, the students must depart the United States and cannot remain through the fall semester. ice.gov/news/releases/sevp-m…





























