Nanda H Krishna
@nandahkrishna.bsky.social
PhD student with @glajoie.bsky.social at Mila – Quebec AI Institute and Université de Montréal. Computational Neuroscience + Deep Learning. Homebrew maintainer, open source enthusiast. Website: https://nandahkrishna.com
2. The non-pretrained POSSM models are on par with baselines such as the GRU. The pre-trained o-POSSM models are statistically better with p < 0.05 (than both non-pretrained POSSM and other baselines).
Hope that answers your questions!
2/2 🧵
Hope that answers your questions!
2/2 🧵
June 8, 2025 at 4:16 PM
2. The non-pretrained POSSM models are on par with baselines such as the GRU. The pre-trained o-POSSM models are statistically better with p < 0.05 (than both non-pretrained POSSM and other baselines).
Hope that answers your questions!
2/2 🧵
Hope that answers your questions!
2/2 🧵
Hi Joao, thank you for the questions! We reported this result as Table 2 in the preprint. To answer:
1. Yes, the first purple bar is trained only on human data while the second is pretrained on monkey data and finetuned on the human data.
1/2 🧵
1. Yes, the first purple bar is trained only on human data while the second is pretrained on monkey data and finetuned on the human data.
1/2 🧵
June 8, 2025 at 4:16 PM
Hi Joao, thank you for the questions! We reported this result as Table 2 in the preprint. To answer:
1. Yes, the first purple bar is trained only on human data while the second is pretrained on monkey data and finetuned on the human data.
1/2 🧵
1. Yes, the first purple bar is trained only on human data while the second is pretrained on monkey data and finetuned on the human data.
1/2 🧵
Late to the thread but here’s the proceedings version from ICLR ‘24 (up to date and much better formatted 😇): arxiv.org/abs/2505.17003
Sufficient conditions for offline reactivation in recurrent neural networks
During periods of quiescence, such as sleep, neural activity in many brain circuits resembles that observed during periods of task engagement. However, the precise conditions under which task-optimize...
arxiv.org
May 23, 2025 at 2:25 AM
Late to the thread but here’s the proceedings version from ICLR ‘24 (up to date and much better formatted 😇): arxiv.org/abs/2505.17003
We’ve uploaded talk recordings to the workshop website (neurofm-workshop.github.io) if you’re interested to check them out!
COSYNE 2025 Workshop - Building a foundation model for the brain
Join us to explore neuro-foundation models. March 31-April 1, 2025 in Mont Tremblant, Canada.
neurofm-workshop.github.io
April 5, 2025 at 12:43 AM
We’ve uploaded talk recordings to the workshop website (neurofm-workshop.github.io) if you’re interested to check them out!
Thanks again to all our speakers, attendees, and the entire team: @averyryoo.bsky.social, @colehurwitz.bsky.social, @mehdiazabou.bsky.social, Eva Dyer, and @tyrellturing.bsky.social for making this happen. 🙌
April 5, 2025 at 12:41 AM
Thanks again to all our speakers, attendees, and the entire team: @averyryoo.bsky.social, @colehurwitz.bsky.social, @mehdiazabou.bsky.social, Eva Dyer, and @tyrellturing.bsky.social for making this happen. 🙌