Shahab Bakhtiari @ShahabBakht Twitter profile

Pinned Tweet

Shahab Bakhtiari

1 year

Our lab 🐌now has a website and a logo: Check out our website to learn more about our approach and research goals. Big thanks to the great artist and neuroscientist @MariaZamfirPhD for helping with the logo.

18

32

288

Last Seen Profiles

@striated_YTs_

@AkinaStarr23239

@camelot1071

@motayb_17

@PortoInsider

@LaurencioJ71088

@Jarkaster

@Swpool22

@MeasuredAccoun1

@mwjb1957578

@servimedenuevo

@CHSPrincipalBR

@RaquelS_xxx

@DSPLab_GT

@trotph

@TheKingofToast2

@bandtrashk

@itbmeshorty

@mattholte

@ichigo_ameee

@MM_PolyTIC

@NFTMerryK

@BradenRay51

@NadAbw68272

@portoquinza

@javierjurado123

@fashiondebates

@kiki_z_t

@IqbalAkhtar

@amirakreet

@stw_pdg

@JakeEisenberg_

@JarrettMelendez

@jandakembangstw

@PearlOtter

@bokeplokalmalam

Shahab Bakhtiari

@ShahabBakht

2 years

🚨News🚨Absolutely thrilled to announce that, starting January 2023, I will be an Assistant Professor in the department of psychology at @UMontreal . I will be developing a NeuroAI research lab focused on visual perception and learning in artificial and biological neural networks

79

29

500

Shahab Bakhtiari

@ShahabBakht

3 years

Check out our new paper: Can a single neural network explain specialized pathways of the visual system? TLDR; yes, but only if you train with a self-supervised predictive loss function! Here is a summary of our paper: (1/n)

The functional specialization of visual cortex emerges from training parallel pathways with...

The visual system of mammals is comprised of parallel, hierarchical specialized pathways. Different pathways are specialized in so far as they use representations that are more suitable for support...

www.biorxiv.org

4

87

337

Shahab Bakhtiari

@ShahabBakht

16 days

I’ll be recruiting 1-2 graduate students for Fall 2025 to work on visual learning generalization in humans and artificial neural networks. If you’re interested, apply! Check out our lab website () and reach out if you need more info. RT please.

snail

We are a NeuroAI laboratory in the Department of Psychology at the Université de Montréal. We aim to understand the algorithmic and neuronal principles underlying animals' visual processing and...

www.snailab.ca

Mila - Institut québécois d'IA

@Mila_Quebec

19 days

📷 Meet our student community! Interested in joining Mila? Our annual supervision request process for admission in the fall of 2025 is starting on October 15, 2024. More information here

0

3

28

5

95

266

Shahab Bakhtiari

@ShahabBakht

3 years

If 16,000 A100 GPUs is the answer, what is the question?

AI at Meta

@AIatMeta

3 years

Meta is announcing the AI Research SuperCluster (RSC), our latest AI supercomputer 💻 for AI research. RSC will allow our researchers to do new, groundbreaking experiments in #AI . Learn more about RSC and the important role it will play:

43

223

919

17

4

215

Shahab Bakhtiari

@ShahabBakht

4 years

Does Tweeting Improve Citations? A randomized trial tl;dr: yes it does.

2

62

201

Shahab Bakhtiari

@ShahabBakht

2 months

Neuroscientists should closely follow the mechanistic interpretability efforts in AI and the debates around it, if they’re not already. Eg see how much the example below (+ the whole thread) is relevant to causal manipulation experiments in the brain.

Aaron Mueller

@amuuueller

3 months

Second big issue: if there are multiple causes of the same effect, it would be very easy to miss both of them if we only ablate one at a time. A related concept is preemption: one cause prevents another from having any effect, and so we miss the preempted cause.

1

2

27

4

26

190

Shahab Bakhtiari

@ShahabBakht

8 months

Hard to be a computational neuroscientist without having been influenced by their works:

Theoretical and computational neuroscientists net 2024 Brain Prize

Larry Abbott, Terrence Sejnowski and Haim Sompolinsky split $1.45 million in recognition of their decades of work uncovering principles of brain structure and function.

www.thetransmitter.org

4

28

160

Shahab Bakhtiari

@ShahabBakht

5 years

No, this isn’t from @tyrell_turing et al recent perspective on @NatureNeuro . This is David Robinson trying to make a similar point in 1992:

8

31

144

Shahab Bakhtiari

@ShahabBakht

4 years

“Here, we demonstrate that predictive coding converges asymptotically (and in practice rapidly) to exact backprop gradients on arbitrary computation graphs using only local learning rule”

Predictive Coding Approximates Backprop along Arbitrary Computation...

The backpropagation of error (backprop) is a powerful algorithm for training machine learning architectures through end-to-end differentiation. Recently it has been shown that backprop in...

openreview.net

2

28

132

Shahab Bakhtiari

@ShahabBakht

3 years

Neuro-AI folks, how do you keep up with both neuro and AI literature? What’s your strategy? It’s a full-time job!!! (And yes, giving up is always an option)

19

11

131

Shahab Bakhtiari

@ShahabBakht

2 years

I very much agree with @markdhumphries that evolution doesn't care what we call a brain area. But, what evolution might care about is modularity as an efficient way of expanding a constantly adapting system. A 🧵:

Mark Humphries

@markdhumphries

2 years

Ah, I see the brain localisation wars have started up again Just going to lob this in and run back to my dugout:

13

56

343

6

22

123

Shahab Bakhtiari

@ShahabBakht

5 years

By far, one of the best analysis I’ve ever read on academia as a business model, the role of postdocs as its horsepower, and how twitter can break down the whole system for good. Can Twitter Save Science? - Alex Danco's Newsletter

Can Twitter Save Science?

Two Truths and a Take, Season 2 Episode 4

danco.substack.com

3

41

113

Shahab Bakhtiari

@ShahabBakht

3 years

Representational drift in the mouse visual cortex; by Daniel Deitch, Alon Rubin, Yaniv Ziv The intrinsic structure of movies representation doesn't change over time despite drift in the coding of the neurons supporting it.

0

23

107

Shahab Bakhtiari

@ShahabBakht

4 years

Octopus arms show photostatic response to light, and most interestingly, the motor response is channeled through the central nervous system (not local reflexes). So it’s kinda like having an 👁 in your palm, and moving your arm based on what your hand “sees”.

3

11

99

Shahab Bakhtiari

@ShahabBakht

4 years

It’s hard to fully understand the importance of her work unless you’re doing science in a developing country where universities and research institutes can’t afford outrageous subscription fees.

RIOT Science Club

@riotscienceclub

4 years

We're still hyped about last week's RIOTS, where we had Alexandra Elbakyan ( @ringo_ring ) share her thoughts on open knowledge, copyright and #SciHub . If you've missed it (or want to rewatch it): 📺 Huge thanks again to @mariiabocharova for translating

5

146

516

0

13

95

Shahab Bakhtiari

@ShahabBakht

26 days

Not all visual features are treated equally in brains or artificial neural networks; some are favored by more neurons. What are the behavioural and learning consequences of these biased representations? I discuss this question in a new blog post: (1/4)

Representational asymmetries and their consequences in neural networks: another shared principle? |...

Neural networks, both biological and artificial, don’t always distribute neurons evenly across input features. Some features may be encoded more frequently by neurons than others. This kind of...

snailab.notion.site

3

12

87

Shahab Bakhtiari

@ShahabBakht

4 years

Is a ‘beach animal’ a computer?

9

8

83

Shahab Bakhtiari

@ShahabBakht

2 years

It’s worth re-reading Olshausen and Field’s 2006 “What is the other 85% of V1 doing?” to see how far we’ve come since then: 1/n

What Is the Other 85 Percent of V1 Doing?

Abstract. The primary visual cortex (area V1) of mammals has been the subject of intense study for at least four decades. This chapter identifies five fund

academic.oup.com

2

11

80

Shahab Bakhtiari

@ShahabBakht

10 months

This work from @FieteGroup looks pretty cool: Showing how structure in the sensory cortex (hierarchy, topography, etc.) can emerge from self-organizing dynamics and spontaneous activity on a cortical sheet. Looking forward to reading this one.

Self-organized emergence of modularity, hierarchy, and mirror reversals from competitive synaptic...

A hallmark of the primate visual system is its architectural organization consisting of multiple distinct (modular) areas that connect hierarchically. These areas exhibit specific spatial organizat...

www.biorxiv.org

1

4

80

Shahab Bakhtiari

@ShahabBakht

8 months

“our results suggest that the ventral visual cortex, like DNN models, provides a texture-like basis set of features, and that further neural computations, perhaps downstream of IT, are necessary to account for the shape selectivity of visual perception.”

Texture bias in primate ventral visual cortex

To accurately recognize objects despite variation in their appearance, humans rely on shape more than other low-level features. This is in contrast to leading deep neural network (DNN) models of...

openreview.net

2

17

73

Shahab Bakhtiari

@ShahabBakht

4 years

I'm learning a lot from this book - by Striedter & Northcutt.

5

9

73

Shahab Bakhtiari

@ShahabBakht

2 years

Prestige seeking behaviors are built into the fabric of academia (and its evaluation systems). Pre-pub peer review & journals have evolved to serve this need in academics. Demolishing it will leave an empty space that will be filled by alternatives that are not necessarily better

8

10

68

Shahab Bakhtiari

@ShahabBakht

1 year

Join our NeuroAI workshop in Montreal 🧠🤖 Check out the list of speakers, register, and submit an abstract.

Guillaume Lajoie

@g_lajoie_

1 year

🚨Announcing a Workshop on advances in NeuroAI🚨: Oct 10-13, held at @Mila_Quebec , Montréal. Abstract submission is open, travel grants for trainees available, speaker list and more 👉 IN-BIC, @ai_unique , @CIFAR_News , @pimsmath , @IVADO_Qc

2

60

142

0

13

68

Shahab Bakhtiari

@ShahabBakht

3 years

Watching #AlphaFold2 unfolding, it’s hard to resist the question of what’d be the neuroscience problem that could be pushed toward solving by throwing engineering and compute resources at it. Some might argue that neuro isn’t even at the stage of having such well defined question

6

64

Shahab Bakhtiari

@ShahabBakht

1 year

Had heard it from others, but now it’s firsthand experience. The joy of hanging out with these young and fun scientists is the best part of running a lab. The first SNAIL social event 🐌

1

2

62

Shahab Bakhtiari

@ShahabBakht

2 years

"Dissociation in neuronal encoding of object versus surface motion in the primate brain" Showing long-range motion direction selectivity in V4 (up to 29% of neurons, using neuropixel recordings)

Dissociation in neuronal encoding of object versus surface motion in the primate brain

Bigelow et al. identify a novel motion computation in the primate ventral stream, complementary to the well-established system in the dorsal visual stream, that is sensitive to long-range motion and...

www.cell.com

0

13

60

Shahab Bakhtiari

@ShahabBakht

4 years

A figure in a recent paper by @talia_konkle and Alvarez () reopened an old question for me. The figure shows that an AlexNet architecture with group normalization has more similar representations to the brain than one with batch normalization.

2

15

57

Shahab Bakhtiari

@ShahabBakht

3 years

Science has definitely become more inclusive with the virtual conferences. Someone who hasn't dealt with visa offices can't see the true point of virtual conferences.

David Cox

@neurobongo

3 years

When live conferences do come back, I do think we should keep virtual participation as an option, to allow broader participation. But I feel like all-virtual conferences are slowly strangling academia.

4

0

27

1

9

53

Shahab Bakhtiari

@ShahabBakht

1 year

Make sure you submit a request if you’re interested in joining the NeuroAI community in Montreal

Mila - Institut québécois d'IA

@Mila_Quebec

1 year

Interested in joining Mila's research community? Our annual supervision request process for new Mila students is starting this Sunday October 15, 2023. For more information ➡️

2

46

149

0

6

55

Shahab Bakhtiari

@ShahabBakht

9 months

New paper @UMontreal @Mila_Quebec (In collab with Chris Pack's lab @TheNeuro_MNI @mcgillu ) We studied Visual Perceptual Learning (VPL) in distinguishing stimuli that have "asymmetric" representations in the visual cortex. How does this asymmetry affect VPL? 🧵(1/10)

Journal of Vision

@ARVOJOV

9 months

Shahab Bakhtiari @ShahabBakht et al. @UMontreal find that asymmetric stimulus representations bias visual perceptual learning.

0

1

14

2

5

53

Shahab Bakhtiari

@ShahabBakht

3 years

mini two-photon microscope for fast, high-res, multiplane ca imaging of over 1,000 neurons at a time in freely moving mice: "A novel technique for successive imaging across multiple, adjacent FOVs enabled recordings from more than 10,000 neurons ..." 🤯

1

12

54

Shahab Bakhtiari

@ShahabBakht

3 years

Interesting, new paper from Cichy lab "The spatiotemporal neural dynamics of object location representations in the human brain" by @MonikaGraumann et al. Object loc emerges in the ventral areas ~ 150 ms later on a high clutter background.

1

7

50

Shahab Bakhtiari

@ShahabBakht

2 years

The next generation of big neuro projects could look something like GPT-JT-6B. A community based, distributed, large-scale model, optimized to align with all the neural/behavioral data that we could put together. What % of the neuro community would get on board with that?

Jim Fan

@DrJimFan

2 years

This enables geo-distributed computing across cities or even countries. Now everyone can BYOC (“Bring Your Own Compute”) and join the training fleet to contribute to the open-source development. The scheduling algorithm makes no assumption about the device types. 3/

4

9

73

7

3

51

Shahab Bakhtiari

@ShahabBakht

2 years

Plenty of gems for new PIs in this note by @behrenstimb : Especially liked this one: “If you ever become Buzsaki, Deisseroth, or Dolan you can have people that work for you. Until then, you work for them.”

1

8

50

Shahab Bakhtiari

@ShahabBakht

1 year

Every time a neuroscientist hints on a podcast that deep learning in neuroscience is merely engineering and regression and not science, one GPU dies.

1

49

Shahab Bakhtiari

@ShahabBakht

7 months

Another cool work on texture / shape bias but in large vision-language models: “we find that VLMs are often more shape-biased than their vision encoders, indicating that visual biases are modulated to some extent through text in multimodal models”

Are Vision Language Models Texture or Shape Biased and Can We Steer Them?

Vision language models (VLMs) have drastically changed the computer vision model landscape in only a few years, opening an exciting array of new applications from zero-shot image classification,...

arxiv.org

Shahab Bakhtiari

@ShahabBakht

8 months

“our results suggest that the ventral visual cortex, like DNN models, provides a texture-like basis set of features, and that further neural computations, perhaps downstream of IT, are necessary to account for the shape selectivity of visual perception.”

2

17

73

2

9

46

Shahab Bakhtiari

@ShahabBakht

3 years

This is very exciting 🎉🥂🥳

Patrick Mineault

@patrickmineault

3 years

Super excited that both @ShahabBakht and I's papers were accepted in NeurIPS as spotlights! Hope you're ready for a whirlwind tour of how the brain solves motion! Tweeprint here 👇

2

5

42

4

1

45

Shahab Bakhtiari

@ShahabBakht

4 years

This is so annoying 😤

じゃがりきん

@jagarikin

4 years

ゴールする瞬間なにかが起きる…！

109

3K

11K

1

10

46

Shahab Bakhtiari

@ShahabBakht

2 months

It’s quite possible that mouse and monkey V1s operate at different processing stages, as suggested in this super interesting review paper by Jon Kaas et al.:

Escaping the nocturnal bottleneck, and the evolution of the dorsal and ventral streams of visual...

Early mammals were small and nocturnal. Their visual systems had regressed and they had poor vision. After the extinction of the dinosaurs 66 mya, some but not all escaped the ‘nocturnal bottleneck’...

royalsocietypublishing.org

Martin Hebart

@martin_hebart

2 months

Another study highlighting that visual findings from mice don't generalize to primates. 2 thoughts: 1) It's fascinating that mice have no orientation columns - architectural constraints? 2) It's crazy nobody has directly shown orientation columns in humans

3

25

126

0

6

45

Shahab Bakhtiari

@ShahabBakht

2 years

I’ll be talking about Self-Supervised Learning and how it can be used for modeling and studying the development of animals’ visual systems: what have been done and what should we do next. Join us if you’re interested.

Researcher

@ResearcherApp

2 years

Join us for a #ResearcherLive session on "Self-supervised learning: A new lens on animal visual development" with @ShahabBakht , @UMontreal @Mila_Quebec 🗓️ Tue, 18 April 2023 🕓 5 pm BST / 4 pm GMT Register now! 👇

0

11

1

3

43

Shahab Bakhtiari

@ShahabBakht

2 years

I'm very happy to have the opportunity of building my lab in Montreal, one of the best and fastest growing NeuroAI ecosystems in the world. I'll be recruiting trainees at all levels. If you're interested in working at the intersection of neuroscience and AI, drop me a message.

3

5

42

Shahab Bakhtiari

@ShahabBakht

2 years

Lastly, to all Iranian students that are affected by the recent events: if you're considering to apply for graduate programs, please let me know if I can be of any help.

2

42

Shahab Bakhtiari

@ShahabBakht

4 years

An interesting read ...

2

5

41

Shahab Bakhtiari

@ShahabBakht

3 years

Eve Marder at @neuromatch 4.0 panel discussion: "If every single postdoc is convinced that he or she should have bona fides in machine learning, what does it tell us about herds?"

7

3

41

Shahab Bakhtiari

@ShahabBakht

3 years

Next time you delete your large model’s checkpoints, beware of the moral ambiguities of your action

Ilya Sutskever

@ilyasut

3 years

it may be that today's large neural networks are slightly conscious

450

562

3K

1

4

39

Shahab Bakhtiari

@ShahabBakht

2 years

I’ve migrated to Neuromatch’s mastodon instance. A great community of neuro and NeuroAI people already there. Join us if you haven’t already. (And for those who wonder if instances matter: yes - local timeline is where you see the difference)

Neuromatch Social

We are a cooperatively-governed instance. Please read our About page for further details.

neuromatch.social

0

9

40

Shahab Bakhtiari

@ShahabBakht

2 years

With prompts in Farsi, #stablediffusion generates a lot of beautiful Islamic architectures and Persian carpets, which have nothing to do with the content of the prompts.

2

0

41

Shahab Bakhtiari

@ShahabBakht

3 years

PaLM explaining jokes to us, DALL-E drawing our dreams, and now learning that transformers have been like the hippocampus all along … what a week!

Jack Hessel

@jmhessel

3 years

"An oil pastel painting of a skeptical researcher absolutely amazed by what he's seeing on a computer" (AI-generated image created by DALL-E with a prompt I wrote, top 1 sample)

7

28

374

2

4

39

Shahab Bakhtiari

@ShahabBakht

4 years

@hardmaru This old video of Gibson’s experiment is relevant. Cliff avoidance existed even ~ 6 months old babies:

1

0

40

Shahab Bakhtiari

@ShahabBakht

3 years

Deep learning studies are getting closer and closer to monkey ephys: spend ~ 6-12 months train the first one, look at the data, if found anything interesting, go for the second one.

Oriol Vinyals

@OriolVinyalsML

3 years

The Deep Learning Devil is in the Details. I love this work from @IrwanBello and collaborators in which they show how training "tricks" improve ~3% absolute accuracy on ImageNet, progress equivalent to years of developments and research! Paper:

16

227

1K

2

3

38

Shahab Bakhtiari

@ShahabBakht

3 years

Number of Vision Science Society (VSS) abstracts with 'neural networks' in title since 2001 Looks like a paradigm shift ... And no sign of saturation! Something happened in 2016 - maybe Tensorflow?! Or lagged response to @dyamins et al (2014)?

2

7

38

Shahab Bakhtiari

@ShahabBakht

4 years

Natural scene representation in mouse V1 requires postnatal visual experience

0

3

31

Shahab Bakhtiari

@ShahabBakht

4 months

Interpreting any failure of AI models as a failure of their "understanding" has proven to be a mistake. Soon, a model will generate accurate gymnastics videos, and we will have to choose between accepting the AI's understanding of human body motion or moving the goalposts.

🌟Cheshire Cat ᓚᘏᗢ,

@autismsupsoc

4 months

CW: Body Horror? This AI video attempt to show gymnastics is one of the best examples I have seen that AI doesn’t actually understand the human body and it’s motion but is just regurgitating available data. (Which appears to be minimal for gymnastics)

447

1K

6K

7

3

34

Shahab Bakhtiari

@ShahabBakht

3 years

@fchollet The hard part is to get your arXiv preprint seen by as many ppl as possible. That’ll have a huge effect on the “adoption” part. One could put brilliant work on arXiv but might not survive the overly skewed attention competition out there, especially given corporates’ PR machinery

1

0

35

Shahab Bakhtiari

@ShahabBakht

8 months

Can’t ignore potential links with low-rank communication subspaces observed in the brain, eg

Feedforward and feedback interactions between visual cortical areas use different population...

Nature Communications - How cortical areas interact via feedforward and feedback signaling remains unclear. Here, the authors recorded from V1 and V2/V4 in macaque visual cortex and found that...

www.nature.com

Andreas Kirsch 🇺🇦

@BlackHC

8 months

2/ GaLore takes a novel approach compared to methods like LoRA and ReLoRA. Instead of doing low-rank projection in weight space (W = W_0 + A @B ) and baking this into the weights every T steps, GaLore performs low-rank projection in gradient space: G = P @ G' @ Q^T.

1

0

24

2

6

34

Shahab Bakhtiari

@ShahabBakht

3 months

Even if you’re NOT a big fan of my research, but have a guess what my next paper would be, I’m very eager to listen.

Kording Lab 🦖

@KordingLab

4 months

Dear (prof first name), I have long been a fan of your research. Here is what your logically next paper would be (without saying it) and I would love to work on it and am ideally prepared for it. Can we talk? Yours, (Your first name)

6

10

104

0

1

33

Shahab Bakhtiari

@ShahabBakht

2 years

I wonder if people with “better” passports are aware of the “passport” effect. Do they consider it in their evaluations? It’s a real thing and it can hugely affect one’s career.

Emtiyaz Khan

@EmtiyazKhan

2 years

If I had known earlier, I could have done this before. They didn't direct me to the appointment page until I finished the documents. Emailed the embassy and got a usual response. F*** this s***. I don't miss physical conferences. They are for people with better passports.

7

90

1

7

34

Shahab Bakhtiari

@ShahabBakht

3 years

Google’s new Pathways project is a great example of what corporate labs are good at: scaling up a simple, promising idea and pushing it to its limits. Small academic labs should be good at developing those simple ideas and demonstrating their potentials.

Introducing Pathways: A next-generation AI architecture

blog.google

2

1

34

Shahab Bakhtiari

@ShahabBakht

5 years

David Sussillo mentioned this old paper by Zipser and Anderson in his talk @neuroAIworkshop at #NeurIPS : It’s perhaps one of the first examples of comparing representations in the brain and artificial neural nets.

2

5

34

Shahab Bakhtiari

@ShahabBakht

4 years

This result by SimCLRv2 (a self-supervised model) is impressive. With only 1% of ImageNet labels, SimCLR2 is as good as the fully supervised ResNet-50.

1

5

34

Shahab Bakhtiari

@ShahabBakht

1 year

I love it every time @AnthropicAI mentions the link between their exciting work and prior work in neuroscience. It's not something you see frequently in AI these days

Anthropic

@AnthropicAI

1 year

Our research builds on work on sparse coding and distributed representations in neuroscience, disentanglement and dictionary learning in machine learning, and compressed sensing in mathematics.

1

6

293

0

1

34

Shahab Bakhtiari

@ShahabBakht

3 years

Our NeurIPS 2021 papers now have their own websites. Check them out for the revised papers, codes, and some visualizations.

Blake Richards

@tyrell_turing

3 years

Our two spotlight papers (by @ShahabBakht and @patrickmineault ) at #NeurIPS2021 showing how ANNs can develop dorsal stream representations now have their websites up (with links to the updated papers and code):

0

27

105

0

4

34

Shahab Bakhtiari

@ShahabBakht

3 years

Enjoyed reading this paper by Yaoda Xu and Maryam Vaziri-Pashkam: Careful analysis of representational similarities between 6 areas of human visual cortex and 14 different ImageNet-trained ANNs. Early areas can be ‘fully’ explained, but not higher areas

Limits to visual representational correspondence between convolutional neural networks and the...

Nature Communications - Convolutional neural networks are increasingly used to model human vision. Here, the authors compare the performance of 14 different CNNs and human fMRI responses to...

www.nature.com

0

6

33

Shahab Bakhtiari

@ShahabBakht

4 years

Yeah... couldn’t be better!

0

33

Shahab Bakhtiari

@ShahabBakht

10 months

On the emergence of specialized streams in neural networks: 1/ specialization can emerge without task-specific optimization. Self-supervised learning, as shown in the recent work by @dfinz et al and our previous work, can lead to emerged specialized streams.

1

6

33

Shahab Bakhtiari

@ShahabBakht

2 years

I know that people who follow me here are not doing that for politics. But this is not politics! Last time this happened, thousands were arrested and hundreds - if not more - were killed on the streets of Iran - A preparation step for another bloodbath. #مهسا_امینی

NetBlocks

@netblocks

2 years

⚠️ Confirmed: Real-time network data show a nation-scale loss of connectivity on MCI (First Mobile), #Iran 's leading mobile operator, and Rightel; the incidents come amid widespread protests over the death of #MahsaAmini 📵 📰 Background:

231

2K

4K

0

13

33

Shahab Bakhtiari

@ShahabBakht

2 years

Looking forward to reading this paper paper

bioRxiv Neuroscience

@biorxiv_neursci

2 years

BCI learning phenomena can be explained by gradient-based optimization #biorxiv_neursci

0

3

22

1

2

31

Shahab Bakhtiari

@ShahabBakht

2 years

These findings from @TimKietzmann lab are among the most exciting I’ve seen recently. Especially, if you’re into predictive coding, make sure you read this paper. I also wrote a preview on why I believe this is a very interesting and important paper:

Energy efficiency as a normative account for predictive coding

In this issue of Patterns, Ali et al. demonstrate that predictive coding emerges in an artificial neural network optimized to be energy efficient. The results offer an explanation for why brains may...

www.cell.com

Tim Kietzmann

@TimKietzmann

2 years

Excited that our paper "Predictive coding is a consequence of energy efficiency in recurrent neural networks" is now out in @Patterns_CP ! Much has changed since preprint, so check it out: Work with the fantastic @hellothere_ali , @nasiryahm , and @marcelge !

3

47

154

2

3

31

Shahab Bakhtiari

@ShahabBakht

4 years

Happy to see our paper finally out. Perceptual improvements caused by Visual Perceptual Learning are highly specific to the training stimulus. In this paper, we showed that the specificity of VPL depends on the complexity of the training stimulus.

1

2

29

Shahab Bakhtiari

@ShahabBakht

3 months

This rule also applies spatially not just temporarily. If you’re a researcher in tech (or any field) and you’re not reading anything five metres away from you (metaphorically speaking), your own work will be forgotten five metres from you.

Peyman Milanfar

@docmilanfar

3 months

If you’re a researcher in tech and you’re not reading anything more than five years old, you can be sure your own work will be forgotten five years from now.

18

52

569

0

3

30

Shahab Bakhtiari

@ShahabBakht

4 years

Organization of LGN in primates, tree shrews, squirrels, and cats. Can I have this for the whole visual system please?! Thanks.

1

5

30

Shahab Bakhtiari

@ShahabBakht

4 years

@neuroecology @DynamicEcology “Neural mechanisms for interacting with a world full of action choices” by Cisek and Kalaska For me, it broke down the perception/action dichotomy by using a neuronal interpretation of Gibsonian affordances.

2

3

30

Shahab Bakhtiari

@ShahabBakht

4 years

Finally ... humans playing Atari during fMRI "Using deep reinforcement learning to reveal how the brain encodes abstract state-space representations in high-dimensional environments"

1

3

29

Shahab Bakhtiari

@ShahabBakht

4 years

I’m old enough to remember invitation-only Gmail. I’m not impressed by your Clubhouse shenanigans kiddo!

1

29

Shahab Bakhtiari

@ShahabBakht

2 years

Day 108 of PI life: This pile of unread papers is gonna crash down and kill me one day (literally and metaphorically). #AcademicChatter

4

0

28

Shahab Bakhtiari

@ShahabBakht

5 years

So, if we regularize an ANN to have similar representations as those of mouse visual cortex, the ANN becomes more robust to adversarial attacks. Here is the most interesting part of the paper for me:

arxiv

@arxiv_org

5 years

Learning From Brains How to Regularize Machines.

0

4

11

0

6

29

Shahab Bakhtiari

@ShahabBakht

2 years

Behavior shapes retinal motion statistics during natural locomotion Another great work from Hayhoe lab! Describing statistical properties of the retinal motion patterns during walking, and how they're affected by gaze location and behavior.

0

9

28

Shahab Bakhtiari

@ShahabBakht

3 years

@neuroecology At least twice according to Striedter & Northcutt

0

1

29

Shahab Bakhtiari

@ShahabBakht

2 years

Iranian girls and their revolution on Vienna’s streets #MahsaAmini #مهسا_امینی

0

5

26

Shahab Bakhtiari

@ShahabBakht

6 months

Me: “what a cool model … we should give it a try in the lab” Model:

3

0

28

Shahab Bakhtiari

@ShahabBakht

3 months

Experiencing major FOMO this year. Will be watching the stream. Enjoy it down there, y’all.

CogCompNeuro

@CogCompNeuro

3 months

For folks who aren't able to attend in-person, we are excited to be able to stream all talks this year on our youtube channel: Papers can be found here: Please share!

0

140

397

0

1

28

Shahab Bakhtiari

@ShahabBakht

4 months

Anyone wondering how ANNs and deep learning have advanced our understanding of the brain should read this paper. I was fortunate to see these results at various stages of their development, and they only got better and better.

Adrien Doerig

@AdrienDoerig

4 months

1/13 Heavily updated preprint! We show that the contextual information encoded in Large Language Models (LLMs) is beneficial for modelling the complex visual information extracted by the brain from natural scenes. 🧵

2

50

164

1

3

28

Shahab Bakhtiari

@ShahabBakht

2 years

Join us in this exciting talk series. I’ll delve into the recent advances in self-supervised learning in AI and their applications in modeling the visual system.

Researcher

@ResearcherApp

2 years

Join us in our upcoming event series, #AI in #Neuroscience , where our experts will explore and share the latest insights. @achterbrain @crntozlu @ShahabBakht Register now! 👇

0

1

6

0

3

27

Shahab Bakhtiari

@ShahabBakht

3 years

GACs have become my favourite way of getting an updated view of a comp neuro/cog field. Whoever came up with the idea deserves the Nobel peace prize. This one from #CCN2021 was great:

CCN 2021: How does visual experience shape representations and...

How does visual experience shape representations and transformations along the ventral stream?Maria Bedny, Nancy Kanwisher, Olivier Collignon, Ilker Yildirim...

www.youtube.com

0

2

27

Shahab Bakhtiari

@ShahabBakht

3 years

As much as I was shocked by the ignorance of the original consciousness tweet, I'm way more shocked by how it's turned my timeline upside down. This level of influence and reach can definitely be put to a better use.

3

0

27

Shahab Bakhtiari

@ShahabBakht

10 months

I believe it’s because neuroscience lacks an easy-to-use metric to measure transformative-ness. There is no list of major open questions that everyone agrees on. We’re divided into small subfields (for better or worse), busy with our own domain-specific questions. (1/2)

Erik Hoel

@erikphoel

10 months

The lack of replies to this reinforces my own belief: neuroscience has come up with many cool results in the past few decades, but none that are truly transformative

5

4

33

2

3

26

Shahab Bakhtiari

@ShahabBakht

3 years

There are so many interesting observations in this paper: The most interesting one imo: 1- "[CLIP] scores close to humans across all of our metrics presented most strikingly in terms of error consistency"

Partial success in closing the gap between human and machine vision

A few years ago, the first CNN surpassed human performance on ImageNet. However, it soon became clear that machines lack robustness on more challenging test cases, a major obstacle towards...

arxiv.org

4

5

27

Shahab Bakhtiari

@ShahabBakht

4 years

Generalization in data-driven models of primary visual cortex "We demonstrate that a representation pre-trained on thousands of neurons from various animals and scans generalizes to neurons from a previously unseen animal" #ICLR2021_submission

0

3

26

Shahab Bakhtiari

@ShahabBakht

4 years

A great demonstration of all the driving skills I had to unlearn before getting my Canadian driving license.

Ali Noorani

@ali_noorani_teh

4 years

Let’s say driving in Tehran is a bit difficult

95

647

2K

1

26

Shahab Bakhtiari

@ShahabBakht

4 years

Half a century ago, neurons firing in real-time

Hubel and Wiesel Cat Experiment

The work of Canadian David Hubel and Swede Torsten Wiesel of the Johns Hopkins University in Baltimore provided clues to how the brain sees images provided b...

www.youtube.com

Elon Musk

@elonmusk

4 years

@Neuro_Skeptic Will show neurons firing in real-time on August 28th. The matrix in the matrix.

179

367

4K

2

9

24

Shahab Bakhtiari

@ShahabBakht

1 year

Beautiful work presenting a generative model of brain connectome! Easy to imagine all the interesting things that can be done with this approach to bring evolution and development into NeuroAI.

Danyal

@DanAkarca

1 year

So happy to share! 🎉 Recently, we’ve seen 🔥 work using generative models to elucidate candidate principles of neural connectivity. Our v2: Inspired by redundancy reduction, our new model can generate both the topology & weights of the connectome:

2

34

114

1

5

26

Shahab Bakhtiari

@ShahabBakht

2 years

I’m amazed by how easily you can break LLMs by including some simple relational queries or info in the prompt. No matter how good they’re in generating high quality texts or images, they easily fail in dealing with relations. And that, imho, says a lot about their shortcomings

2

1

26

Shahab Bakhtiari

@ShahabBakht

4 years

Imo, one of the most influential works in computational (vision) neuroscience. It was my first encounter with an optimization-based approach to visual coding which was a tempting alternative to the more common tuning-curve-ology approach

Gabriel Peyré

@gabrielpeyre

4 years

Oldies but goldies: B. Olshausen, D. Field, Emergence of simple-cell receptive field properties by learning a sparse code for natural images, 1996. Showed that dictionary learning on natural image produces wavelet-like atoms.

1

42

236

1

25

Shahab Bakhtiari

@ShahabBakht

4 months

I haven't seen any satire account being taken so seriously so often. Not sure what that says about their jokes or the subjects of their jokes.

5

0

25

Shahab Bakhtiari

@ShahabBakht

2 years

How AI transformers mimic parts of the brain

0

6

25

Shahab Bakhtiari

@ShahabBakht

3 years

This dataset is 🤯 From the preprint () We all need to actively think what can be done with this dataset.

Allen Institute

@AllenInstitute

3 years

Announcing the most detailed examination of mammalian brain circuitry to date. This data release marries a 3D wiring diagram with the function of tens of thousands of neurons. @bcmhouston @PrincetonNeuro @IARPAnews @awscloud @googlecloud 🧵 👇

3

252

677

1

5

25

Shahab Bakhtiari

@ShahabBakht

3 years

A very useful study, with many practical insights about self-supervised learning (SimCLR here) Trying to answer important questions about e.g. the effect of dataset size, domain etc on SSL

3

5

25

Shahab Bakhtiari

@ShahabBakht

2 years

Scaling is important, but I’ll refer everyone who believes scaling is _all_ you need to Chen’s beautiful work, as an example of how far we can go with genuine ideas, without scaling.

Chen Sun 🧠🤖🇨🇦

@ChenSun92

2 years

(1/n) Hello Friends! I wanted to share my paper with a #tweeprint . We made an RL module we call contrastive introspection (ConSpec). It enables learning when rewards are sparse and multiple key steps are required for success, as often occurs in real life

10

45

179

1

5

25

Shahab Bakhtiari

@ShahabBakht

2 years

@NicoleCRust The optimist inside me would argue that building brains will lead to understanding brains (à la Feynman) and eventually fixing them, so the current fast rate of progress in building brains is actually part of the progress toward fixing them.

0

23