Danyal Akarca (@DanAkarca) / X

Danyal Akarca

2,841 posts

Danyal Akarca

@DanAkarca

Co-founder @CallosumAI | Machine intelligence, brains & chips | prev Schmidt fellow @imperialcollege, @schmidtfutures, @cambridge_uni.

London, England

Joined January 2012

Pinned
Danyal Akarca
@DanAkarca
Feb 26
Today we're launching @CallosumAI out of stealth. I’m incredibly excited. We think AI & compute is about to change fundamentally. The current paradigm scaled up on an assumption that single models on identical chips would be enough. That shaped everything - models, data
Callosum
@CallosumAI
Feb 26
The future of compute & AI is heterogeneous. That future is being built by Callosum. Today we launch with new breakthroughs made possible by heterogeneity, new scaling principles & a roadmap reimagining compute & AI infrastructure: callosum.com/blog/introduci…
Exclusive: Callosum, a startup looking to break Nvidia’s stranglehold on AI workloads, raises...
From fortune.com
20K
Danyal Akarca
@DanAkarca
Nov 1, 2022
aaaaand that’s the PhD submitted ✅ been helluva ride
Danyal Akarca
@DanAkarca
Jul 9, 2021
My first PhD paper is out in @NatureComms! 🎉 “A generative network model of neurodevelopmental diversity in structural brain organization”, with the great @DuncanAstle @petra_vertes @edbullmore & @calmcbu! Supported by @mrccbu @Cambridge_Trust & @JSMF nature.com/articles/s4146…
Danyal Akarca
@DanAkarca
Dec 21, 2023
We're just getting started in discovering how biology constrains the emergence of intelligence & behaviour, and leveraging this with AI. Spatially embedded neuromorphic networks | @NatMachIntell nature.com/articles/s4225… Thank you to @misicbata + @FilipMilisav on the piece!
12K
Danyal Akarca
@DanAkarca
Jun 27, 2023
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: biorxiv.org/content/10.110…
00:00
35K
Danyal Akarca
@DanAkarca
Mar 1, 2024
Beyond stoked to finally announce that this summer I’ll be moving to @imperialcollege as an Imperial College Research Fellow supported by @SchmidtFutures!💥 I’ll be working closely at the intersection of Neuroscience & AI w/ a focus on efficiency, topology & information theory.
11K
Danyal Akarca
@DanAkarca
Mar 15, 2024
*Fully funded PhD position opening*💥 I have an opening for my first PhD student to come to @imperialcollege, co-supervised with @neuralreckoning, working on new projects in neuro-AI 🧠 Must be UK home fees status (3 yrs in UK prior). Plz share! Further details below 👇👇
22K
Danyal Akarca
@DanAkarca
Mar 26, 2024
Can't recommend @rasbt's short lectures enough on building a solid intuition behind how self-attention & multi-head attention in transformers work... w/ RNNs (great intro): youtube.com/watch?v=mDZil9… Scaled attention: youtube.com/watch?v=0PjHri… Multi-head: youtube.com/watch?v=A1eUVx…
13K
Danyal Akarca
@DanAkarca
Jan 7, 2023
Viva done! ✅ 🎉 Massive thank you to @DaniSBassett & @rai_bethlehem for examining me. I loved the discussions & look forward to future work in generative models & embedded neural networks. Thank you to my brilliant supervisor @DuncanAstle, colleagues & department @mrccbu!
12K
Danyal Akarca
@DanAkarca
Jun 13, 2022
Just shy of 9 years and 4 degrees later, I’m starting my first full day NOT as a student or medic. So happy to start as a post-doc today at Cambridge, @MRCCBU continuing with @DuncanAstle + the fantastic team! Can’t wait to share our cool new stuff and develop exciting ideas.
Danyal Akarca
@DanAkarca
Sep 13, 2023
Finally on my way down under! ✈️ 🇦🇺 Extremely excited to be visiting the @jmacshine group, working closely with @chrisjackwhyte on biophysical generative models of spiking networks, with planned visits to @AFornito @AndrewZalesky @ParkvilleGeek. Hmu if you’d like to meet! 👋
6.4K
Danyal Akarca
@DanAkarca
Nov 21, 2022
💥So excited to finally share this project that @achterbrain and I have co-led!💥 What happens to RNNs if you give them some brain-inspired constraints of 3D structure and neural communication? Turns out, a lot! We call these networks spatially-embedded RNNs (seRNNs).
bioRxiv Neuroscience
@biorxiv_neursci
Nov 18, 2022
Spatially-embedded recurrent neural networks reveal widespread links between structural and functional neuroscience findings biorxiv.org/cgi/content/sh… #biorxiv_neursci
Danyal Akarca
@DanAkarca
Jun 1, 2023
Where computational/AI models meet brain development: historical work, recent advances, across species & scales - with many new & exciting horizons! See our recent review in @TrendsCognSci w/ @DuncanAstle & Mark Johnson. 🧵👇
Duncan Astle
@DuncanAstle
Jun 1, 2023
V. excited by this in @TrendsCognSci. A long time in the writing, w/ Mark Johnson and @DanAkarca. Can computational models tell us something fundamental about cognitive and brain development? We think so. We term it "computational neuroconstructivism". doi.org/10.1016/j.tics…
9.3K
Danyal Akarca
@DanAkarca
Apr 30, 2024
One of the most beautiful papers I've seen on the topic of dynamical systems & complex networks. TLDR: Understanding underlying dynamics + constraints gives a much richer (+ accurate) view of function on graphs. From @GZamora_Lopez & Matthieu Gilson
pubs.aip.org
An integrative dynamical perspective for graph theory and the analysis of complex networks
Built upon the shoulders of graph theory, the field of complex networks has become a central tool for studying real systems across various fields of research. R
6.1K