Kording —-& Lab 🦖 @KordingLab Twitter profile

Last Seen Profiles

@sulenursa

@izs

@StarInCluster

@xyletherin

@TAKAHIGASHINO

@smugmarius

@lynnsdeepsocial

@CentralCl20

@b4lmalaga

@RobinAllersOslo

@rainmaker1980

@JimTurner_OL

@RRizzuti10

@comf0rtcrwd

@MasjidHidaayah

@sodoghoul

@manarma27227015

@ihavedisease

@zachary01306057

@jayden_rolle

@Ego_GG_

@swedtree

@KeoshaMoon

@yanoselena5

@HumanDisaster44

@jenrubysie

@kathy_diamonds

@itsasalax

@growbananahigh

@JonEKummelAtely

@8keeemo

@SquadUpEsp

@IcaroSouza57643

@sfccmoBB

@matsh_

@QueerNerdyLefty

Kording —-& Lab 🦖

@KordingLab

4 years

Spectacularly happy @neuromatch news: we just received an emergency license from the US Government to operate in Iran. Thanks to the community that helped make this happen! ❤️ We can now push further towards the dream of global science without borders.

66

397

2K

Kording —-& Lab 🦖

@KordingLab

6 years

Science is about overcoming failures. List: (1) once I got to be professor, my first 23 grant proposals we're rejected. I am very comfortable now. (2) I was mediocre at school (~B) and particularly bad at stats. Much of my lab's work is close to stats now.

21

311

1K

Kording —-& Lab 🦖

@KordingLab

11 months

Woo hoo. I got to hold a copy of our book in my hands today! Great feeling

15

102

984

Kording —-& Lab 🦖

@KordingLab

2 years

Just my yearly reminder: there is a standard format behind (probably) most published papers. I highly recommend all students look at that format: makes writing easier for my students (and myself)

Ten simple rules for structuring papers

journals.plos.org

6

245

931

Kording —-& Lab 🦖

@KordingLab

4 years

All materials from @neuromatch academy on one page now: If you teach data science (in any subdomain) you got to look at this. Amazing work by our volunteer team.

12

316

833

Kording —-& Lab 🦖

@KordingLab

2 years

Massive whitepaper just dropped on why neuroscience progress should continue to drive AI progress: Argues for an embodied turing test. Needed: real interdisciplinary people, shared platform(s), fundamental research

28

194

743

Kording —-& Lab 🦖

@KordingLab

4 years

Today is my birthday. 🎁 🎂 And I want to use it as an opportunity to thank all the people who work to make neuroscience better, more fun, and more human. 🎉 ❤️ I feel blessed to be part of this community.

78

7

692

Kording —-& Lab 🦖

@KordingLab

1 year

Machine learning can easily produce false positives when the test set is wrongly used. Just et al in @NatureHumBehav suggested that ML can identify suicidal ideation extremely well from fMRI and we were skeptical. Today retraction and our analysis of what went wrong came out.

15

222

678

Kording —-& Lab 🦖

@KordingLab

4 years

I have a causality joke, but not sure it is the reason why you laugh.

Venkat Ramaswamy

@VenkRamaswamy

4 years

I have a theory joke, but not sure if it would work in practice.

14

32

459

48

59

673

Kording —-& Lab 🦖

@KordingLab

3 years

Wrote a paper don't know where to send it? Got a paper and don't know who should review it? Worried you miss related papers? All scientists should use jane for such questions: Been in love with it for years and only just realized not everyone knows.

15

181

650

Kording —-& Lab 🦖

@KordingLab

5 years

Just re-advertising "ten simple rules for structuring papers" - it helps my trainees write far better drafts. And do I wish that the papers I review followed it. #readability

Ten simple rules for structuring papers

journals.plos.org

5

256

621

Kording —-& Lab 🦖

@KordingLab

4 years

Neuroscientists: if you want to learn computational techniques apply to our online summer school: Please help reach out to the community that is not on twitter and language communities. This will be great.

The Neuromatch Academy - Student Application Round 1

What is The Neuromatch Academy? The Neuromatch Academy will be an open, online, 3-week intensive tutorial-based computational neuroscience training event (July 13-31, 2020). Participants from...

docs.google.com

35

339

597

Kording —-& Lab 🦖

@KordingLab

2 years

What is the biggest question you think neuroscience should answer in the next decade?

213

96

570

Kording —-& Lab 🦖

@KordingLab

1 year

I know I publish too many papers. But some of my colleagues publish 100 papers a year: e.g. and there are assistant professors who publish >50. Seems widespread () Is this a good development? How do people even do that?

Thousands of scientists publish a paper every five days

Nature - To highlight uncertain norms in authorship, John P. A. Ioannidis, Richard Klavans and Kevin W. Boyack identified the most prolific scientists of recent years.

www.nature.com

52

116

521

Kording —-& Lab 🦖

@KordingLab

2 years

Hugely honored and challenge accepted. Not an April’s fools joke. Starting my tenure today.

CIFAR

@CIFAR_News

2 years

Congratulations to Konrad Kording @KordingLab , new co-director of the CIFAR Learning in Machines & Brains program alongside Yoshua Bengio. Our deep thanks to @ylecun for his long tenure in this role as he moves to program advisor.

4

15

173

69

17

527

Kording —-& Lab 🦖

@KordingLab

4 years

I consistently learn more preparing teaching materials than doing research. Just saying.

9

42

500

Kording —-& Lab 🦖

@KordingLab

1 year

Neurons can use their spiking to identify their causal effects, allowing an approximation to gradient descent learning. Finally accepted

Neural spiking for causal inference and learning

Author summary Despite significant research, models of spiking neural networks still lag behind artificial neural networks in terms of performance in machine learning and modeling cognitive tasks....

journals.plos.org

8

86

456

Kording —-& Lab 🦖

@KordingLab

1 year

Aligning latent representations of neural activity [Review] With @evadyer we ask how recordings across times, conditions or animals can be aligned. No supervised signals needed! Try the codes here:

6

110

446

Kording —-& Lab 🦖

@KordingLab

4 years

So for 4 months I have pretty much worked on @neuromatch academy all the time. Every day I woke up, found new problems to solve. And helped solve them. It feels so weird to wake up and not do that. Want to again express appreciation for team!

12

437

Kording —-& Lab 🦖

@KordingLab

5 years

Our opinion piece that deep learning should guide neuroscience research just came out:

A deep learning framework for neuroscience

Nature Neuroscience - A deep network is best understood in terms of components used to design it—objective functions, architecture and learning rules—rather than unit-by-unit...

www.nature.com

5

178

437

Kording —-& Lab 🦖

@KordingLab

4 years

Inviting all neuroscientists. Join us for zoom drinks tomorrow (saturday) evening at 9pm east (EDT). I hear there are breakout rooms (and otherwise we can realtime break out). Will send a link tomorrow.

Kording —-& Lab 🦖

@KordingLab

4 years

It is time to #CancelEverythingNow . But only the offline things. Why isn't online-drinking-together a thing yet?

14

9

126

30

89

439

Kording —-& Lab 🦖

@KordingLab

2 years

AI and natural intelligence (NI) appear so different. A group of us asked “why?” The answer: we have to go back to the future! A thread. 1/6

10

107

432

Kording —-& Lab 🦖

@KordingLab

5 years

What does it mean to understand a neural network: - Tim Lillicrap and me try to think deeply about how pondering neural networks changes our goals as neuroscientists.

What does it mean to understand a neural network?

We can define a neural network that can learn to recognize objects in less than 100 lines of code. However, after training, it is characterized by millions of weights that contain the knowledge...

arxiv.org

15

134

434

Kording —-& Lab 🦖

@KordingLab

4 years

Thanks to all 2862 #neuromatch2020 participants, speakers, helpers, team, and matchers. Love you all. Was so much fun. Lets get together again sometimes soon. Talks (invited and contributed) are all recorded - simply go to .

25

75

430

Kording —-& Lab 🦖

@KordingLab

2 years

I learned: Understanding Neuro is useless for ai. Understanding Ai is useless for Neuro. Understanding Neuro is useless for Neuro. And understanding ai is useless for ai. ;) wonderful threads all around.

26

27

397

Kording —-& Lab 🦖

@KordingLab

4 years

At NMA ( @neuromatch ) we are devastated about our inability to serve the Iranian community. We want to build a political outreach powerhouse (e.g. do interviews with journals, op-eds, twitter campaigns). Who can join us make this be effective?

47

112

392

Kording —-& Lab 🦖

@KordingLab

4 years

We asked what people want to learn at summer school. Here are the most popular ones: ML, DL, Bayes, Behavior, RL, Stats, Networks, RNNs, fMRI, Dynamics, Dimensionality rediction, Causality, Model comparison, spikes, Information Theory. + many flavors of "good modeling" Thank you

27

36

390

Kording —-& Lab 🦖

@KordingLab

4 years

Are you a neuro/ psych/ data grad student or postdoc who needs a remote position in summer/fall 2020? Visa depends on it? Reluctant to ask? Let us try to help. NeuromatchCareers will explore potential matches that match your interests and skill set.

13

245

383

Kording —-& Lab 🦖

@KordingLab

4 years

Big open summer school (July) in computational neuroscience. What would you like your students to learn? We want to run the best course possible for the field. Help us by filling out the topics-to-learn-for-your-students questionnaire:

What do you want your students to learn at a computational neuroscience summer school

We are planning a large online computational neuroscience summer school in July 2020, a summer school where all neuroscientists should be welcome and that would be entirely free. Prerequisite would...

docs.google.com

18

153

378

Kording —-& Lab 🦖

@KordingLab

2 years

There are now so many notebooks / Jupyter books that are awesome at explaining stuff. Way better than papers. Why is there not a place to find them all? Why are they not cross linked? It’s a new ecosystem. It’s not papers with code. It’s mostly learning with code. I prefer.

18

33

373

Kording —-& Lab 🦖

@KordingLab

3 years

After living here for more than a decade, I became a citizen of the United States of America. Honored to join such a high energy place. I hope I will contribute to making it a little bit nicer. 🇺🇸❤️

18

3

375

Kording —-& Lab 🦖

@KordingLab

5 years

It is important to realize that authors, reviewers, and editors are human beings. So. We wrote this paper on quasiexperimental causality which just appeared. - Somewhere it mentioned that people who are rich "own yachts" are probably happier.

8

103

371

Kording —-& Lab 🦖

@KordingLab

3 years

Know python and data science but want to add deep learning? Not just the superficial ideas but intuition and implementation? Intensive summer school. Team based. 3 weeks in August: - amazing lecturers, the best tutorials I can imagine.

11

113

356

Kording —-& Lab 🦖

@KordingLab

1 year

The big logical gap within systems neuroscience Neuroscientists make causal statements about brains. “We want to understand how the brain works, how it computes, how to fix it.” The relevant causal problems are high dimensional with billions of neurons nonlinearly affecting one…

50

56

344

Kording —-& Lab 🦖

@KordingLab

2 years

Paper's official summary "having a million subjects is way better than just thousand for brain imaging based phenotype prediction" - real message "brain data is practically useless for phenotype prediction, even if we had millions of subjects". Time to shut down the field?

Danilo Bzdok

@danilobzdok

2 years

Our next piece on performance scaling #deeplearning in imaging neuroscience: Marc’s work projects 3-9 fold improvements in prediction performance for behavioral and mental health phenotypes when moving from one thousand to 1 million samples @uk_biobank

6

44

138

24

64

340

Kording —-& Lab 🦖

@KordingLab

2 years

Are papers that raise doubt about the core of a field essentially impossible to publish or are they not written? Who has experience with those? Feel free to DM if publicly discussing is not good. What examples of such papers are there?

75

58

337

Kording —-& Lab 🦖

@KordingLab

5 years

Reviewers were amazingly helpful but one saw an opportunity to have some fun pointing out "Mo money mo problems":

2

54

334

Kording —-& Lab 🦖

@KordingLab

4 years

Why is it ok to introduce someone as "got his PhD at xxx"? Isn't it elitist? Why does it even matter decades beyond this PhD? Why is it considered more important than what they studied? Or their contributions?

46

21

326

Kording —-& Lab 🦖

@KordingLab

4 years

Causality is a central, arguably *the* central topic in neuroscience. Why do I have trouble finding more than a handful of people worldwide who actually work on causal inference?

47

33

324

Kording —-& Lab 🦖

@KordingLab

3 years

No. No neuralink chips will probably never replace language. A) technically we are far further away. B) there is simply no reason/market/point to it. Humans can already talk faster than they think.

Neuralink Brain Chip Will End Language in Five to 10 Years, Elon Musk Says

Elon Musk stated that the development of Neuralink will allow humans to communicate using a universal language in the near future. But the priority of the brain chip will focus first on solving...

www.sciencetimes.com

27

41

318

Kording —-& Lab 🦖

@KordingLab

3 years

So if @elonmusk does on monkeys what countless people (a dozen labs?) did decades ago on both monkeys and humans everyone is super impressed and it is national news. Why?

39

47

320

Kording —-& Lab 🦖

@KordingLab

3 years

Academia should be a joint voyage of discovery.

Leonhard Schilbach

@leoschilbach

3 years

4

82

440

7

34

317

Kording —-& Lab 🦖

@KordingLab

3 years

3 week deep learning summer school (Aug 2-20) by at @neuromatch is coming. Will have projects. And obviously tutorials where people learn the skills by doing them. Anything under- or overemphasized in this syllabus relative to what is most useful?

23

64

308

Kording —-& Lab 🦖

@KordingLab

5 years

Every neuroscientist and ml scientist should read Rich Sutton's "bitter lesson" assay.

17

113

302

Kording —-& Lab 🦖

@KordingLab

4 years

And for people just tuning in. We will run a massive, open for all neuroscientists, free or nominal price, summer school that will basically teach all these things in July (3 weeks). All timezones. Tight work with TAs.

7

53

306

Kording —-& Lab 🦖

@KordingLab

3 years

Why I believe that most things we learn about functional connectivity are not helpful at understanding how brains work:

The lure of misleading causal statements in functional...

As neuroscientists we want to understand how causal interactions or mechanisms within the brain give rise to perception, cognition, and behavior. It is typical to estimate interaction effects from...

arxiv.org

18

64

306

Kording —-& Lab 🦖

@KordingLab

3 years

Want to learn computational neuroscience? Not just the superficial ideas but all the implementations? Want to learn how to implement it all? 3 weeks in July: @neuromatch here: - amazing lecturers, the best tutorials I can imagine.

6

124

300

Kording —-& Lab 🦖

@KordingLab

4 years

I think that the brain almost certainly approximates gradient descent. And here is why: Any learning episode only appears to change the brain a tiny bit. 1/5

36

78

297

Kording —-& Lab 🦖

@KordingLab

5 years

Now, the editor did spot this. And did what editors do with references - make them better. Altogether this was a great experience with @NatureHumBehav @marike_cogneuro et al.

2

18

296

Kording —-& Lab 🦖

@KordingLab

4 years

How to run big (neuro)science conferences online — what we learned. Let us make online conferences awesome.

11

97

285

Kording —-& Lab 🦖

@KordingLab

5 years

Officially announcing NBDT (), a new journal by the data, behavior, physiology, theory community for this community. We tested it in open beta and just published the very first issue. Officially accepting submissions.

On the Subspace Invariance of Population Responses | Published in Neurons, Behavior, Data analysis,...

By Elaine Tring, Dario Ringach. The responses of neurons in cat visual cortex to sums of gratings reveals a symmetry: subspace invariance

nbdt.scholasticahq.com

8

141

283

Kording —-& Lab 🦖

@KordingLab

2 years

Hot take: Proofs in many ICML papers are a thin veneer of mathematical sophistication, meant to show the authors' refinement. They are not generally meant to contribute to math or to apply to reality but as proving membership in a tribe. A bit like an NFT.

Thomas Steinke

@shortstein

2 years

I reviewed 3 papers for ICML 2022 @icmlconf . All 3 had fundamental flaws (like theorems that are demonstrably false). None of the other 6 reviewers noticed these flaws and gave 👍. This is the sorry state of ML conferences...

46

122

1K

20

22

278

Kording —-& Lab 🦖

@KordingLab

5 years

If people think that computer simulations can replace animals any time soon they are sorely mistaken. Big mistake by the EPA: And I am saying that as a leading brain simulation scientist.

18

130

276

Kording —-& Lab 🦖

@KordingLab

4 years

This account is, obviously, apolitical, as it should be. But it is not a political question if this country should have justice for all. Science is a collaborative endeavor, and injustice against anyone threatens everyone and poisons the trust that underlies science.

6

25

272

Kording —-& Lab 🦖

@KordingLab

5 months

"Why the simplest explanation isn’t always the best" - commentary with @evadyer highlighting how dimensionality reduction does not usually give us what we want. "Major Achievement: Dino scatterplot in paper" unlocked.

Why the simplest explanation isn’t always the best | Proceedings of the National Academy of Sciences

Why the simplest explanation isn’t always the best

www.pnas.org

7

71

267

Kording —-& Lab 🦖

@KordingLab

2 years

I am incredibly inspired to be named (as of today) the "Nathan Francis Mossell University Professor" (). He founded Douglass Memorial Hospital. Above all, a man of principles. I don't think I can ever live up to his standards but I'll try. 1/2

28

5

273

Kording —-& Lab 🦖

@KordingLab

4 years

Big open summer school (July) in computational neuroscience. What would you like to learn? We want to run the best course possible for the people who would participate. Help us by filling out the topics-to-learn questionnaire:

What do you want to learn at a computational neuroscience summer school

We are planning a large online computational neuroscience summer school in July 2020, a summer school where all neuroscientists should be welcome and that would be entirely free or extremely cheap....

docs.google.com

6

111

267

Kording —-& Lab 🦖

@KordingLab

4 years

Neuromatch reached 4k registrants shortly after closing. Massive worldwide audience. Great matching, debates, talks, posters etc. Amazing community. Global. Positive. Watch *all* talks/debates/etc here for free:

5

53

263

Kording —-& Lab 🦖

@KordingLab

11 months

Landmark story shows that brain activity is spatiotemporally low-pass. I am shocked!

Geometric constraints on human brain function

Nature - Cortical and subcortical activity can be parsimoniously understood as resulting from excitations of fundamental, resonant modes of the brain’s geometry rather than from modes of...

www.nature.com

12

47

257

Kording —-& Lab 🦖

@KordingLab

2 years

Was just going to share that, as of today, U2 is taking over @neuromatch for a payment of $213M in cash and an undisclosed sum in stock. The leadership will be immediately replaced by a group of MBAs with the main goal of lowering payment of TAs. Today is a truly special day.

25

11

259

Kording —-& Lab 🦖

@KordingLab

10 months

Meh. Mandatory code and data sharing. Rigorous science standards. Ways of appealing exploitative practices and abuse. Outlaw paywalls. Efficient reviewer time use. Post publication review. Mandatory preprints. Literally everything is more important than first author stuff.

Patrick Hsu

@pdhsu

10 months

The most impactful near-term thing we can do in science publishing to improve research culture is to prominently credit and display co-first authors. It should be immediately clear. Right now, readers have to click on author contributions or decode weird superscript symbols 👇

34

39

347

10

36

255

Kording —-& Lab 🦖

@KordingLab

2 years

10 years. 1 text book. Interested in Bayesian models of perception and action? Help us with feedback

Wei Ji Ma

@weijima01

2 years

Bayesian modeling of perception and action: draft of book with Konrad Kording ( @KordingLab ) and Daniel Goldreich ( @dangoldreich ) done after 10 years... To be published by @mitpress but already freely available on . Comments are welcome (link there)!! 1/2

20

298

1K

6

37

253

Kording —-& Lab 🦖

@KordingLab

3 years

Yearly reminder. I recommend to all new students of science our guide breaking down the standard structure of papers. Gotta know the structure, to creatively break it.

Ten simple rules for structuring papers

journals.plos.org

3

77

253

Kording —-& Lab 🦖

@KordingLab

3 years

Opinion: every student needs to learn how to code. Whatever we do, we live in a world shaped by code. Understanding code is necessary to shape it. We need to teach it so well that every student wants to learn it. Enable improving the world. Think about what that means.

John Butler

@jslbutler

3 years

Replace the word neuroscience with any discipline from science to the arts. Learning to code is essential. It teaches algorithmic logical thinking which helps break problems down.

2

5

56

11

35

251

Kording —-& Lab 🦖

@KordingLab

4 years

ML folks: what are the best ideas from neuroscience that you found actually useful? because "neurons exist" is so 1890s. And "attention helps" is so 1890s as well. #NeurIPS2019

21

40

249

Kording —-& Lab 🦖

@KordingLab

1 year

Who of the famous people in the field would be ready to publicly talk about their non-rigorous research in the past? I think that few people have ever published 10 papers without major rigor issues in at least one of them. Let's normalize talking about rigor.

29

45

248

Kording —-& Lab 🦖

@KordingLab

4 years

Ila fiete nicely shows that, even if we recorded all neurons, inferring connectivity may not be possible

Systematic errors in connectivity inferred from activity in strongly recurrent networks

Nature Neuroscience - The authors demonstrate that strongly recurrent circuits inferred from neural activity, even with unlimited data from every neuron, are biased. Synapses are inferred between...

www.nature.com

12

75

240

Kording —-& Lab 🦖

@KordingLab

3 years

I guess that all that staying at home made me blue. But then, I guess I should have chosen the @neuromatch color.

7

4

239

Kording —-& Lab 🦖

@KordingLab

1 year

So what was cool about the @neuralink presentation was not any specific thing they presented. It's rather that they do many of the coolthings that us neuroscientists have been doing. And all of them a tad better than the field did.

16

25

239

Kording —-& Lab 🦖

@KordingLab

3 years

Looks like I missed it. On google trends deep learning/ neural networks are no longer that hot. Have we passed peak DL?

27

24

240

Kording —-& Lab 🦖

@KordingLab

1 year

So what went wrong? The authors apparently used the test data to select features. Obvious mistake. A reminder for everyone into ML: never use the test set for *anything* but testing. Only practical way to do so in medicine? Lock away the test set till algorithm is registered.

8

33

237

Kording —-& Lab 🦖

@KordingLab

2 years

Human Psychophysics on Bayesian Behavior was my life for well more than a decade. Now @JABrantl finds it in baseball on millions of pitches. Linear response to perturbations, manipulations of priors, human Behavior appears Bayesian.

Justin Brantley

@JABrantl

2 years

What do you get when you cross professional baseball ⚾️with Bayes ? Bayesball! I am super excited to announce a new paper out with @KordingLab , titled "Bayesball: Bayesian Integration in Professional Baseball Batters"

11

50

206

3

58

232

Kording —-& Lab 🦖

@KordingLab

4 years

We are now hard at work re-incorporating our Iranian students and TAs before classes begin in a few hours! We will reach out to you soon if you are an Iranian participant. We are sorry for all the stress, but we are so happy with the final outcome.

4

16

232

Kording —-& Lab 🦖

@KordingLab

4 months

Where are all the neural manifold folks at? I was expecting major disagreement with our paper on dimensionality reduction not revealing deep insights.

Why the simplest explanation isn’t always the best | Proceedings of the National Academy of Sciences

Why the simplest explanation isn’t always the best

www.pnas.org

16

33

230

Kording —-& Lab 🦖

@KordingLab

4 years

Just got sent the printed preview version of "Dive into deep learning" for which there is the free interactive online version: - omg this text book is so awesome.

6

55

229

Kording —-& Lab 🦖

@KordingLab

1 year

Ok this intro to neuroscience course is off the charts good. @bingbrunton is setting a new standard. I would be curious what the tech stack you used for making these. So amazing!

Bing Wen Brunton

@bingbrunton

1 year

The only thing I love more than neuroscience is... teaching neuroscience. Here's a video series of my undergrad “Intro to Neuro” class. Adding new topics regularly. Check them out!

49

391

2K

4

28

228

Kording —-& Lab 🦖

@KordingLab

3 years

@RogersLabUCD My first 23 grant proposals got rejected in a row. Just saying. 3 years no grants. don't listen to their bs.

3

6

228

Kording —-& Lab 🦖

@KordingLab

6 years

The brain estimates the probabilities of what you might do before you can even ponder what you will do:

8

119

228

Kording —-& Lab 🦖

@KordingLab

4 years

It seems that the number of simultaneously recorded neurons now grows a little faster than exponential (courtesy Ian Stevenson lab). I welcome the incoming industry innovations!

6

49

222

Kording —-& Lab 🦖

@KordingLab

3 years

There should be a meeting called Neuroangst. All you do is give a talk about why you worry about your past papers’ conclusions. No judgment. Or discussions. Everyone just is supportive and says that you are a great scientist just for being there.

5

16

223

Kording —-& Lab 🦖

@KordingLab

2 years

Surprise guest in the lab today ( @alfcnz ). Had great fun talking about deep learning, teaching, and energy based models.

11

6

219

Kording —-& Lab 🦖

@KordingLab

2 years

Was just going to say that @BlackInNeuro is an amazing organization. Among other things they run a great list (that every Neuro conference organizer should know about) and mentorship programs. And also, they have great merch!

3

14

222

Kording —-& Lab 🦖

@KordingLab

4 years

Today’s #NMAVolunteer is Athena Akrami( @AthenaAkrami ). Athena is a brilliant scientist and leads the Learning, Inference, and Memory lab () who use theory and experiment to ... 1/3

4

13

215

Kording —-& Lab 🦖

@KordingLab

3 years

At @penn I want to start an accessible data science for everyone course. What are the best comparable courses on that?

28

24

212

Kording —-& Lab 🦖

@KordingLab

5 years

The brain is obviously not a Bayes Machine. And DL brains are so 2008. The brain is a blockchain:

15

57

210

Kording —-& Lab 🦖

@KordingLab

2 years

In our exciting new preprint we solve one of the oldest problems of neuroscience.

@[email protected]

@j6m8

2 years

@chazfirestone @KordingLab thrilled to share an early sneak peek at our upcoming preprint:

8

7

70

22

13

209

Kording —-& Lab 🦖

@KordingLab

5 years

unknown to most outside the field, DL often does not outperform far simpler models. With obvious exceptions for NLP, vision, and RL.

Thomas Yeo

@bttyeo

5 years

“This suggests that DL - in fact - does not improve prediction compared to simpler baseline models.“ - controversy about deep learning paper predicting earthquake published in Nature.

3

39

156

9

41

209

Kording —-& Lab 🦖

@KordingLab

7 years

What goes where in a paper. Writing is easy with the standard structure. Finally out:

6

120

207

Kording —-& Lab 🦖

@KordingLab

2 years

Which neuroscience groups/ departments/ centers are cool enough to hire philosophers. Asking for a friend.

35

20

200

Kording —-& Lab 🦖

@KordingLab

5 years

Psych and Neuro stack exchange are finally out of beta. Let us popularize it. I promise that I will answer at least ten questions there this year.

Psychology & Neuroscience Stack Exchange

Q&A for practitioners, researchers, and students in cognitive science, psychology, neuroscience, and psychiatry

psychology.stackexchange.com

2

103

203

Kording —-& Lab 🦖

@KordingLab

5 years

I am in awe of this functional connectivity paper that finally says what everyone always felt: we do it because we want counterfactual causality: - obviously that is still probably impossible but I applaud the honesty.

Advancing functional connectivity research from association to causation

Nature Neuroscience - Many studies focus on neural associations yet understanding the brain will ultimately depend on discovering the causal interactions underlying its functionality. Moving from...

www.nature.com

7

52

203

Kording —-& Lab 🦖

@KordingLab

5 years

I hate having to tell my trainees that from postdoc to tenure they should pretend to be (mostly) one-trick-ponies. But the incentive structure still favors that. More papers -> less training -> more early career success ( -> unfortunately potentially less actual impact)

19

29

203

Kording —-& Lab 🦖

@KordingLab

5 months

So I am in San Francisco, hanging out at a meeting with many startup folks. I love the optimism (which I share). But is it just me or is there a lot of tolerance to not deep dive before trying to move? Why is it not worth to deeply think through ideas before executing on them?

41

8

203

Kording —-& Lab 🦖

@KordingLab

2 years

Within psychophysics, Bayesian models are often great at describing behavior. We wrote a textbook on that. Please give us feedback. Free online. Forever. Will be printed by MIT press from next October afaik. Apologies for tweeting again, we want to get all the feedback we can.

Wei Ji Ma

@weijima01

2 years

Bayesian modeling of perception and action: draft of book with Konrad Kording ( @KordingLab ) and Daniel Goldreich ( @dangoldreich ) done after 10 years... To be published by @mitpress but already freely available on . Comments are welcome (link there)!! 1/2

20

298

1K

7

52

200

Kording —-& Lab 🦖

@KordingLab

4 years

So we are super busy producing tutorials and workflows for @neuromatch academy. Just for people to see how we are experimenting, see the current lectures accompanying the tutorials for Bayes Day here: Tech team is helping us produce very cute videos.

Bayes Day (2020)

Bayes Day, a full day tutorial on Bayes' theorem and its application to neuroscience.

www.youtube.com

5

39

201

Kording —-& Lab 🦖

@KordingLab

5 years

A public service announcement: if you write papers you *must* reveal limitations. (1) anything less makes you a bad scientist. (2) reviewers give you credit (3) you avoid future shame. I recommend at least three paragraphs of limitations. Your paper has less? #YouFoundUnicorn

Kording —-& Lab 🦖

@KordingLab

5 years

@JCashaback A limitations paragraph (or better 3) should be part of every paper. I am trying my best. Seriously. Might even get you published as reviewers hate hidden weaknesses.

0

2

27

5

29

198

Kording —-& Lab 🦖

@KordingLab

5 years

Neuroscience needs causality. And quasiexperiments are a crucial tool. Learn about them in our new review. with top-economist @mioana

Quasi-experimental causality in neuroscience and behavioural research

Nature Human Behaviour - How to establish causal links is a central question across scientific disciplines. Marinescu and colleagues describe methods from empirical economics and how they could be...

www.nature.com

11

69

198

Kording —-& Lab 🦖

@KordingLab

3 years

You learned deep learning and are ready to help others learn it too? @neuromatch may run a large super didactic DL summer school (3 weeks, group based). with lots of awesome people. Let us know if you may be able to TA (paid) with us this summer.

12

83

194

Kording —-& Lab 🦖

@KordingLab

5 years

Our paper "what does it mean to understand a neural network?" is deeply shaping the way I think about the future of neuroscience. Where should we publish it?

Real Neuroscience journal

147

Review journal

133

Machine learning outlet

86

Just show me

427

15

39

189

Kording —-& Lab 🦖

@KordingLab

3 years

My amazing wife just got tenure. So deserved. Such wonderful research. All the things that make the world better.

7

0

189