Profile picture
, 7 tweets, 9 min read
My Authors
Read all threads
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI You don't actually keep up with the neuroscience literature, do you? That has been evident in these conversations... Here, lemme give you a few examples:
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI 1) ANNs optimised on relevant tasks match the representations in human (and primate) cortical areas better than other models developed to date:

journals.plos.org/ploscompbiol/a…

sciencedirect.com/science/articl…

journals.plos.org/ploscompbiol/a…
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI 2) ANNs trained on motor tasks successfully predict both motor behaviour and the distribution of representations in motor cortex:

sciencedirect.com/science/articl…

biorxiv.org/content/biorxi…
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI 3) Recurrent ANNs can model prefrontal cortical activity well, and explain both behaviour and representations related to multi-task learning:

nature.com/articles/s4159…

nature.com/articles/s4159…
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI 4) In my own work, I found that the original predictions of complimentary learning theory from ANNs (psycnet.apa.org/record/1995-42…) predicts the impact of memory consolidation on mouse navigation behaviour:

nature.com/articles/nn.37…
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI And there's more! But, I have to go to lab meeting... Nonetheless, I would encourage you to actually approach the role of ANNs in neuroscience with an open mind. A lot has been accomplished, and there's more to come.
@GaryMarcus @r_chavarriaga @KordingLab @DeepMindAI Here's a couple of reviews for you if you're actually interested and not so ideological that you are incapable of approaching it neutrally (you have not convinced me of this to date):

nature.com/articles/s4159…

biorxiv.org/content/10.110…
Missing some Tweet in this thread? You can try to force a refresh.

Enjoying this thread?

Keep Current with Blake Richards

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

More from @tyrell_turing see all

Profile picture
Blake Richards
@tyrell_turing
1/ SciTwitter: I'm very excited to share our new Perspective article out in Nature Neuroscience today!

nature.com/articles/s4159…
2/ In this piece, we argue that neuroscience would benefit from adopting a framework that parallels the approach to designing intelligent systems used in deep learning.
3/ ANN researchers do not attempt to design specific computations by-hand, instead, they design three core components: (1) architectures, (2) objective functions (cost/loss), and (3) learning rules that provide good inductive biases for learning how to do specific computations.
Read 17 tweets
Profile picture
Blake Richards
@tyrell_turing
1/ New tweeprint from my lab! This one is work done by the amazing @guerguiev, and was inspired by the work of @benlansdell and @KordingLab (who was also a collaborator in this project).

arxiv.org/abs/1910.01689
2/ Our focus in this paper is the question of weight alignment. If you try to work out how to efficiently estimate cost fn gradients in biological networks, you will find that it would be great if feedforward pathways and feedback pathways had symmetric synaptic weights.
3/ But, symmetric synaptic weights is not something we can reasonably assume for biological neural networks, since it implies a magic transmission of weights between different pathways. This is known as the "weight transport problem".
Read 15 tweets
Profile picture
Blake Richards
@tyrell_turing
1/ #tweeprint time everybody! It's about neural coding (and I mean that literally). We asked the following Q: if info is encoded in the neocortex with both rate and synchrony of spikes, do different subtypes of neurons display differential sensitivity to these two info streams?
2/ In the example image above, a binary signal is encoded with either a rate or synchrony code. The rate code uses high rate = 1, low rate = 0. In contrast, in the synchrony code the cells have a constant rate-of-fire, but high synch = 1, low synch = 0.
3/ We hypothesized that PV+ and SST+ interneurons would be differentially sensitive to these two codes, due to their intrinsic and synaptic properties. For example, the short-term depressing synapses of PV+ cells may make them more sensitive to synchrony.
Read 19 tweets

Related threads

Profile picture
Nathan Brouwer
@lobrowR
A basic guide to trouble shooting problems with #rstats
(A thread for my students and others new to R)

Something in R not working? Weird error message?
Go through this list of steps to try to resolve the problem. [thread; suggestions & other tips highly welcome]
@DimperioJane
1) Did you run a line of code without a ")" at the end?

Look at your code in the console. Is there a little "+" at the far left? R is probably waiting for a complete line of code. Click into the console, press ESC, add the ")", and try again
2) Is R waiting for user input?
Occasionally R will ask for user input via a pop up Window or within the console itself. Look in the console see if you need to give a numeric response (option 1, option 2) or look for the GUI's popup.
Read 21 tweets
Profile picture
Alex Zelasi
@AZelasi
I don't know what a hot take is but maybe hot take I guess: most allistic attempts to teach autistic people "social skills" are just attempts to teach us THEIR social skills, with little to no reciprocal effort on their part to learn about autistic social skills.
Also possibly a hot take: autistic and allistic people are each better at certain types of socialising and each group can be said to have specific social skills. Neither group is superior to the other, despite the relentless portrayal of autistic people as having "deficits".
Someone please explain what a hot take is: any allistic person who doesn't acknowledge that autistic people DO have social skills - just different ones - and that autistics are actually, on the whole, better at some forms of communication, is ableist.
Read 35 tweets
Profile picture
foone
@Foone
random thing that annoys me and probably shouldn't:
Googling for a DOS game and getting a suggestion for "$GAMENAME ROM"

WITH VERY FEW EXCEPTION, DOS GAMES DO NOT HAVE ROMS
consoles like the NES have "roms" because the game stores all the game data on a ROM chip.
These chips are called that because they're Read Only Memory: ROM. The data on them is fixed at time of manufacture, by actually encoding it into the structure of the silicon
if you took the ROM chip, chopped off the top, and zoomed way in, you'd find that the silicon looks like this.
There's a bunch of little transistors arranged so that some of them encode 1s and some encode 0, and there's some circuitry to determine which part is read out.
Read 74 tweets
Profile picture
Ayla Ferrone
@AylaFerrone
Have just received transcripts from yesterday's proceedings for the trial against #NXIVM leader Keith Raniere. It's difficult to keep up on the transcripts as they are hundreds of pages long, so be patient. Here's some notes from yesterday where Lauren Salzman testified. #thread
Salzman remained on direct examination. I'm told she likely will remain on the stand through tomorrow morning. Yesterday's trial began with attorneys discussing use of certain email evidence with the judge. Judge bars the exhibits due to specific accusations against Raniere.
Salzman confirming again, as her testimony begins, that she was a first-line master in DOS. She says this means she was enrolled directly under Raniere and that he was her master. She says aside from him, everyone else in DOS was female.
Read 35 tweets
Profile picture
5'7'' Black Male
@absurdistwords
Actually, no. That's not what the data says at all. But I really liked your casual slide into racism.

Stylish AF.
Here. Lemme give you this before the routine starts. Get back to me after you're done!

twitter.com/i/moments/1000…
Actually, this too. I don't wanna repeat myself.

Read 3 tweets

Trending hashtags

#jeffreyepstien #Q1 #DarkToLight #Actually #neuroscience #NeuroThursday #SoMe4Surgery #DiversityJedi #röstamermänskligt #SCBWI #SoMe4Trauma #WakeUpAmerica #PAIN #ImpeachmentHearing #colorectalsurgery #edneuro #actually #physics #NXIVM #Treason #BanFrackingBC #education #DECLAS #Pedonetwork #PainIsComing
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!