Excited to launch Principia, a nonprofit research organisation at the intersection of deep learning theory and AI safety.
Our goal is to develop theory for modern machine learning that can help us understand network behaviors, including those critical for AI safety.
1
Thanks to amazing co-authors William Dorrell, @nishpathead and supervisor @SaxeLab, and a special shoutout to @TPVogels!
Paper: https://t.co/hBGTub36Wa
Find us at NeurIPS, Thur 4:30 pm #2115! We know networks have to be both plastic and stable but we're used to thinking about computations, such as memory, as additional requirements. Instead, we find that almost all stable & plastic networks display simple memory abilities.
We derive toy models that explain these seemingly accidental memories as input-dependent byproducts of movement in the manifold of stable weights - a form of memory - and a nice conceptual insight from meta-learned plasticity rules!
Come to our poster II-068 today to see our latest progress on meta-learning families of plasticity rules in spiking networks, work with @pramesh95 @ppjgpncalves @jakhmack and @TPVogels
Dear #neurotwitter, we dropped this onto @biorxiv_neursci for your perusal. It's an unusual piece from my lab & I really, really enjoyed working on it with @chc1987. Here it is: Metabolically spikes serve neuronal energy homeostasis (and protect neurons). https://t.co/muKt2jNPyG
Congrats @basile_cfx, on your first preprint (and spotlight at #neurips2020) w/ @ejagnes@hisspikeness , @countzerozzz & myself, on how to (re)discover plasticity rules in neuronal networks, by way of smartly applied brute force. (Tweeprint to follow) https://t.co/y9jDVVwnsr
#cosyne2020 is starting soon! Today I'll be presenting a poster about the work I've done with Andrea Luppi and @TPVogels (I-41). Also, check out the work by @basile_cfx, @hisspikeness, @TPVogels and myself, today as well (I-13).