Ryan Young
@synapticsage.bsky.social
Neural network interested in cat videos, brains, AI, and math. Probably in that order. Ph.D., Neuroscience.
Brandeis, Jadhav Lab alumni
Sr ML Engineer @ Mayflower
Julia, Python, Mojo, C++ programmer
Twitter refugee
Brandeis, Jadhav Lab alumni
Sr ML Engineer @ Mayflower
Julia, Python, Mojo, C++ programmer
Twitter refugee
Reposted by Ryan Young
Excited to share that our paper is now published in Nature! In this study, we uncovered a grid-cell-based circuit in medial entorhinal cortex that probes the surrounding environment with theta-paced, sweeping spatial representations 🧵
www.nature.com/articles/s41...
www.nature.com/articles/s41...
February 4, 2025 at 9:12 PM
Excited to share that our paper is now published in Nature! In this study, we uncovered a grid-cell-based circuit in medial entorhinal cortex that probes the surrounding environment with theta-paced, sweeping spatial representations 🧵
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Reposted by Ryan Young
𝗧𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺 𝘄𝗶𝘁𝗵 𝗔𝗜 𝗮𝗻𝗱 𝗿𝗲𝗹𝗮𝘁𝗲𝗱 𝘄𝗮𝘆𝘀 𝗼𝗳 𝘂𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 𝗯𝗿𝗮𝗶𝗻 𝗮𝗻𝗱 𝗯𝗲𝗵𝗮𝘃𝗶𝗼𝗿: Lack of robustness to a qualitatively unknown future in an open world.
"Evolution and The Knightian Blindspot of Machine Learning".
Old lesson about AI from the 1970s that has been forgotten.
arxiv.org/abs/2501.13075
"Evolution and The Knightian Blindspot of Machine Learning".
Old lesson about AI from the 1970s that has been forgotten.
arxiv.org/abs/2501.13075
Evolution and The Knightian Blindspot of Machine Learning
This paper claims that machine learning (ML) largely overlooks an important facet of general intelligence: robustness to a qualitatively unknown future in an open world. Such robustness relates to Kni...
arxiv.org
January 24, 2025 at 10:55 PM
𝗧𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺 𝘄𝗶𝘁𝗵 𝗔𝗜 𝗮𝗻𝗱 𝗿𝗲𝗹𝗮𝘁𝗲𝗱 𝘄𝗮𝘆𝘀 𝗼𝗳 𝘂𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 𝗯𝗿𝗮𝗶𝗻 𝗮𝗻𝗱 𝗯𝗲𝗵𝗮𝘃𝗶𝗼𝗿: Lack of robustness to a qualitatively unknown future in an open world.
"Evolution and The Knightian Blindspot of Machine Learning".
Old lesson about AI from the 1970s that has been forgotten.
arxiv.org/abs/2501.13075
"Evolution and The Knightian Blindspot of Machine Learning".
Old lesson about AI from the 1970s that has been forgotten.
arxiv.org/abs/2501.13075
Reposted by Ryan Young
Check out our #NeurIPS2024 spotlight "Get rich quick: exact solutions reveal how unbalanced initializations promote rapid feature learning."
Thu 12 Dec 4:30-7:30 pm
East Exhibit Hall A-C #2102
Paper: arxiv.org/abs/2406.06158
Thread: x.com/KuninDaniel/...
Thu 12 Dec 4:30-7:30 pm
East Exhibit Hall A-C #2102
Paper: arxiv.org/abs/2406.06158
Thread: x.com/KuninDaniel/...
Get rich quick: exact solutions reveal how unbalanced initializations promote rapid feature learning
While the impressive performance of modern neural networks is often attributed to their capacity to efficiently extract task-relevant features from data, the mechanisms underlying this rich feature le...
arxiv.org
December 12, 2024 at 12:24 AM
Check out our #NeurIPS2024 spotlight "Get rich quick: exact solutions reveal how unbalanced initializations promote rapid feature learning."
Thu 12 Dec 4:30-7:30 pm
East Exhibit Hall A-C #2102
Paper: arxiv.org/abs/2406.06158
Thread: x.com/KuninDaniel/...
Thu 12 Dec 4:30-7:30 pm
East Exhibit Hall A-C #2102
Paper: arxiv.org/abs/2406.06158
Thread: x.com/KuninDaniel/...
Reposted by Ryan Young
New paper out in #Neuron: A general theory of sequential working memory in prefrontal cortex and RNN/SSMs with their exact neural mechanism. Plus unifying this new mechanism with the alternate mechanism of hippocampal cognitive maps! (1/9)
www.cell.com/neuron/fullt...
www.cell.com/neuron/fullt...
December 8, 2024 at 3:27 PM
New paper out in #Neuron: A general theory of sequential working memory in prefrontal cortex and RNN/SSMs with their exact neural mechanism. Plus unifying this new mechanism with the alternate mechanism of hippocampal cognitive maps! (1/9)
www.cell.com/neuron/fullt...
www.cell.com/neuron/fullt...
Reposted by Ryan Young
Discovery of a biomass foam for potentially 98%+ removal of microplastics from water @ScienceAdvances www.science.org/doi/10.1126/...
Revivable self-assembled supramolecular biomass fibrous framework for efficient microplastic removal
Chitin-cellulose fibrous foam constructed by hydrogen bonding–induced self-assembly achieved efficient removal of microplastics.
www.science.org
November 29, 2024 at 7:12 PM
Discovery of a biomass foam for potentially 98%+ removal of microplastics from water @ScienceAdvances www.science.org/doi/10.1126/...
Reposted by Ryan Young
Could we get away with (1) region specific RPEs and (2) no end-to-end training? Is that how the brain works?
My guess is yes to (1), no to (2).
This paper from our "cousins" at @vectorinst.bsky.social is exploring thes issues and has some promising results!
arxiv.org/abs/2411.03604
#NeuroAI 🧪
My guess is yes to (1), no to (2).
This paper from our "cousins" at @vectorinst.bsky.social is exploring thes issues and has some promising results!
arxiv.org/abs/2411.03604
#NeuroAI 🧪
Temporal-Difference Learning Using Distributed Error Signals
A computational problem in biological reward-based learning is how credit assignment is performed in the nucleus accumbens (NAc). Much research suggests that NAc dopamine encodes temporal-difference (...
arxiv.org
November 29, 2024 at 4:32 PM
Could we get away with (1) region specific RPEs and (2) no end-to-end training? Is that how the brain works?
My guess is yes to (1), no to (2).
This paper from our "cousins" at @vectorinst.bsky.social is exploring thes issues and has some promising results!
arxiv.org/abs/2411.03604
#NeuroAI 🧪
My guess is yes to (1), no to (2).
This paper from our "cousins" at @vectorinst.bsky.social is exploring thes issues and has some promising results!
arxiv.org/abs/2411.03604
#NeuroAI 🧪
Reposted by Ryan Young
So, I keep seeing media assertions that the miracle of 2023 — rapid disinflation with solid growth — was something economists considered impossible. Ahem. Here's a figure from Krugman and Wells 7th edition, forthcoming 1/
January 6, 2024 at 5:39 PM
So, I keep seeing media assertions that the miracle of 2023 — rapid disinflation with solid growth — was something economists considered impossible. Ahem. Here's a figure from Krugman and Wells 7th edition, forthcoming 1/
Reposted by Ryan Young
The Neuron as a Direct Data-Driven Controller https://www.biorxiv.org/content/10.1101/2024.01.02.573843v1
The Neuron as a Direct Data-Driven Controller https://www.biorxiv.org/content/10.1101/2024.01.02.573843v1
In the quest to model neuronal function amidst gaps in physiological data, a promising strategy is t
www.biorxiv.org
January 3, 2024 at 3:15 PM
The Neuron as a Direct Data-Driven Controller https://www.biorxiv.org/content/10.1101/2024.01.02.573843v1