Redwood Center for Theoretical Neuroscience
@Redwood_Neuro
Followers
1K
Following
40
Statuses
37
The Redwood Center for Theoretical Neuroscience is a group of faculty, postdocs and graduate students at UC Berkeley. Account run by students.
Berkeley, CA
Joined January 2018
New work published by Adrianne Zhong in the Redwood Center's DeWeese group! 🥳🥖
Very excited to share our new paper out in PRL today (as an Editors’ Suggestion)! We unify thermodynamic geometry 🔥 and optimal transport 🥖 providing a coherent picture for optimal minimal-work protocols for arbitrary driving strengths 💪 and speeds ⏱️ !
0
0
1
Thank you for visiting!
My recent talk at UC Berkeley: Neurons as feedback controllers. Thanks Bruno Olshausen for inviting me to the terrific Redwood Institute for Computational Neuroscience via @internetarchive
0
0
0
Bruno Olshausen will be speaking tomorrow (Monday) at 1:30 PDT: "Neural computations for geometric reasoning". Abstract:
Join us next week at our intelligence workshop @SimonsInstitute! Schedule: Register online for both in-person and streaming. FANTASTIC lineup of speakers: 1/3
0
0
6
RT @neur_reps: NeurIPS is here! Come check out the NeurReps Workshop on Sat Dec 3, ft. invited talks by ⭐️ @TacoCohen ⭐️ Irina Higgins @Dee…
0
16
0
RT @TrentonBricken: Super super excited for my first day as a visiting researcher at @Redwood_Neuro. Will be here and living in the Bay are…
0
2
0
Congrats! We’re very excited about this workshop @NeurIPSConf
Our workshop on Symmetry and Geometry in Neural Representations has been accepted to @NeurIPSConf 2022! We've put together a lineup of incredible speakers and panelists from 🧠 neuroscience, 🤖 geometric deep learning, and 🌐 geometric statistics.
0
0
7
If you haven't already, scooch on over to our Seminars page to check out a recent talk by Jeff Hawkins and Subutai Ahmad on the Thousand Brains theory of intelligence:
0
0
8
Exciting new research from recent RCTN grads @DylanPaiton, @charles_irl, and others on how neuroscience-inspired recurrent layers for image processing with sparse coding grant increased selectivity and robustness to adversarial inputs
Hey twitter, check out my new paper! We investigate how competition through network interactions can lead to increased adversarial robustness and selectivity to preferred stimulus. paper: explainer 🧵: 1/8
0
0
3
RT @charles_irl: new paper with @DylanPaiton and others in @Redwood_Neuro now out in @ARVOJOV -- connecting ideas in visual neuroscience wi…
0
2
0
...resisted the urge to add thug lyfe sunglasses. Can correct later if necessary
0
0
1
Preprint spotlight: Charles Frye (@charles_irl) et al. demonstrate a surprising weakness in critical point-finding methods used to support the "no bad local minima" hypothesis re: neural net losses. See🧵for a quick explainer, or dive deep🏊♀️in the paper:
There's a theory out there that neural networks are easy to train because their loss f'n is "nice": no bad local minima. Recent work has cast doubt on this claim on analytical grounds. In new work, we critique the numerical evidence for this claim. 🧵⤵
0
3
9
A special day in the Redwood Center as we host the thesis seminar of Yubei Chen!
0
0
7
Our weekly CompNeuro journal club is back up and running for the quarter. Today, we discussed Cheung et al’s preprint "Superposition of many models into one" from @berkeley_ai and @Redwood_Neuro. Led by AJ Kruse and Satpreet Singh.
0
0
3
Congratulations to Sylvia and Spencer who passed their qualifying exams this week! 🎉🎊
0
0
8
We are looking forward to hearing from our latest seminar speaker, @pkdouglas16, who is coming up from UCLA! Her talk on modeling latency and structural organization of connections in the brain is at noon tomorrow.
1
0
3
New preprint spotlight: Saeed Saremi and Aapo Hyvarinen develop a framework for density estimation and sampling based on a unification of kernel density and Empirical Bayes. You can read all about it on the arXiv:
0
4
17
New preprint spotlight: Louis Kang and Mike DeWeese present an attractor network model of grid cells which exhibits replay and theta sequences. Sound interesting? Learn more here!
0
1
6
New preprint spotlight: Brian Cheung (@thisismyhat), Alex Terekhov, Yubei Chen (@Yubei_Chen), Pulkit Agrawal (@pulkitology) and Bruno Olshausen report a simple but effective technique for building neural networks that can solve many different tasks:
0
6
14
New preprint spotlight: Charles Frye (@charles_irl), Neha Wadia, Mike DeWeese, and Kris Bouchard tell us about how to estimate the critical points of a deep linear autoencoder, complete with gotchas, tricks, and how this might be extended to other models.
0
3
8