Manta Array @ArrayManta

research engineer, computational neuroscience, neuromechanics ⚾ follows not equal to endorse. Peace to all. master Jack ass of some trades, PhD of none Joined November 2015

Tweets

3K
Followers

365
Following

1K
Likes

8K

Earl K. Miller @MillerLabMIT

19 hours ago

New results! The prefrontal cortex uses a low-dimensional “coding space”. Info can be captured in just 3–6 dimensions. The brain compresses what we remember. Low-dimensional prefrontal representations of objects during working memory doi.org/10.64898/2026.… #neuroscience

13 60 299 18K 199

View Details

Manta Array @ArrayManta

20 hours ago

@ChrisWLynn fwiw nice paper, I work on very similar models. Abstracts can be challenging to compress both neurobio and neuroIPs perspectives and assumptions. Your stat mech perspective is important and best to your research program

0 0 2 58 0

View Details

Earl K. Miller @MillerLabMIT

23 hours ago

The cortex is not best understood as a set of fully separate, self-contained modules. Instead, regions work together in a more integrated, network-like way. Brain-wide distributed processing underlying natural vision and audition doi.org/10.64898/2026.… #neuroscience

5 24 115 9K 46

View Details

Vineet Tiruvadi, MD PhD @vineettiruvadi

2 days ago

A latent space vs THE latent space.

0 1 2 136 0

View Details

Bernardo Sabatini @blsabatini

4 days ago

1/n Are biological neurons linear-nonlinear computers, like perceptrons, or is their output governed by non-linear interactions between inputs? If the activity of a neuron is well fit by linear models that sum inputs, does that mean that the neural computation actually is linear?

3 21 135 28K 113

View Details

Manta Array @ArrayManta

2 days ago

@karlrohe Yes! RMT and PCA for the people!

0 0 2 43 0

View Details

Manta Array @ArrayManta

2 days ago

@karlrohe when statistically evaluating power laws in eigenspectra of large random matrices, is it important to consider eigenvalues/vector are not iid when applying this: arxiv.org/abs/0706.1062

1 0 1 62 0

View Details

nature @Nature

4 days ago

Labs are teaming up to run a host of huge projects that aim to bring rigour to cognitive science. go.nature.com/4x4Ia2z

3 13 40 14K 12

View Details

Manta Array @ArrayManta

3 days ago

@PessoaBrain @_fernando_rosas @KordingLab I think your point about time binning is important too and actually related to what Konrad is getting through coarse graining effects

0 0 3 74 0

View Details

Manta Array @ArrayManta

3 days ago

@ChoppinBravos Maybe he had a bet with Ozzie Albies

0 0 0 313 0

View Details

Manta Array @ArrayManta

2 weeks ago

Map and territory thing i think. An actual individual biological neuron is the territory for many neuroscientists and it is known to be a more complex entity, universal function approximation is a formal feature of ANNs

Brad Aimone @jbimaknee

2 weeks ago

I feel like this paper is simply stating something we know, which is that any unknown neuron function can be approximated by a circuit of simple neurons. Doesn't this fall directly out of universal function approximation and threshold gate theory?

1 0 1 595 0

1 1 3 243 0

View Details

Manta Array @ArrayManta

6 days ago

@kanair maybe can be on the safe side and effectively treat them as they somewhat are at least valued as if they are, humans naturally do this even with machines like old automobiles

0 0 4 76 0

View Details

bayes @bayeslord

6 days ago

I’m guilty of this to be clear but man techbro philosophy is mostly quite bad. “First principles thinking” but also you forgot to read up on the hundreds or thousands of years of prior thought on the subject

24 25 355 14K 58

View Details

Tyler Austin Harper @Tyler_A_Harper

a week ago

If the entire tenure and promotion system weren't predicated on churning out a bunch of (mostly?) slop that no one reads, then there would be no incentive to author a bunch of fake AI papers. Academic culture was broken before AI, the chatbots just made this impossible to ignore.

6 17 126 15K 12

View Details

François Chollet @fchollet

a week ago

Einstein on (not) using NL for invention: "The words or the language, as they are written or spoken, do not seem to play any role in my mechanism of thought"

70 51 558 58K 161

View Details

Manta Array @ArrayManta

a week ago

@TrackingActions @TonyZador @CSHL tells the truth

0 0 1 161 0

View Details

Manta Array @ArrayManta

a week ago

The Math is not the Territory: Navigating AI Consciousness Discourse Not so different and still important @bayesianboy

Maxwell Ramstead @mjdramstead

a week ago

My two cents on AI consciousness discourse: Folks need to learn the distinction between computational modeling (using computational tools to model the dynamics of a real system, which need not be computational) and the computational theory of mind (cognition is computation)

4 9 26 4K 7

0 1 7 1K 4

View Details

NASA Administrator Jared Isaacman @NASAAdmin

2 weeks ago

The near impossible is becoming possible. We are building toward a sustained human presence at the lunar South Pole. It begins with Phase 1: CLPS landers and LTV rovers testing the “science of survival” on the lunar surface before heavy HLS cargo landers deliver the mass and infrastructure needed for an enduring presence. We are building the Moon Base for all we will learn, the innovation that will improve life on Earth, the inspiration for the next generation of explorers, and to master the skills needed for where we will inevitably go next...Mars. The Golden Age of lunar exploration has begun.

583 2K 11K 2.4M 559

View Details

Manta Array @ArrayManta

2 weeks ago

Breaking: A Bayesian brain explains another Bayesian brain as just being Bayesian

Alex Imas @alexolegimas

2 weeks ago

I think this is just being Bayesian. People use AI in writing for two things mainly: 1) improving their writing after idea generation and 2) as a shortcut to idea generation (eg write me a few paragraphs making argument X and provide support). Most people are likely doing both