🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Thanks Scientific American for featuring our work! scientificamerican.com/article/new-to…

loss go down

My first first-author paper: AtMan has been accepted at NeurIPS Conference🥳 It's a SOTA method to trace the outputs of an LLM back to the input space (gradcam pro max without the gradients) arxiv.org/abs/2301.08110

rise and shine boys

this is still relevant

🚀At NeurIPS Conference tomorrow? Don't miss Nikolas McNeal and Mainak Deb 's poster at the UniReps workshop on the adversarial sensitivity of vision encoding models of fMRI responses! A brief teaser about what they find.

Unlike the brain, most ANNs lack any kind of organization of units🧠 In our paper (ICLR 2025) with mayukh09 Apurva Ratan Murty, we present TopoLoss, a simple, scalable way to add brain-like topography to AI models with minimal performance loss🧵 toponets.github.io

Very proud to share our ICLR 2026 paper: TopoNets! High-performing vision and language models with brain-like topography! Expertly led by grad student mayukh09 and Mainak Deb! A brief thread..

signal boosting this because it is both intuitive and extremely cool

I always love to see brain topology inspired work. The brain saves so much on compute with useful topology/wiring already between retina and the first layer of the visual cortex. We should not forget where neural networks came from and should keep imitating nature. Great work!

Watch how the science-selective neurons self-organize in a layer when I train a GPT w/ topoloss toponets.github.io This figure could not make it to the paper but I liked it

Mainak Deb mayukh09 Apurva Ratan Murty Bringing brain-like topography to ANNs with minimal performance loss sounds like a huge step toward more efficient and structured AI. Excited to see how TopoLoss impacts future architectures!

My work got featured in TEDAI Vienna👀 AtMan (NeurIPS 2023) explains LLM outputs, telling you which part of the context was critical. ted.com/talks/jonas_an… neurips.cc/virtual/2023/p…

Come check out our Spotlight paper - TopoNets at #ICLR2025! DM me if you're interested in discussing structured LLMs, vision models and brain-inspired algorithms in Singapore! 24 Apr 3:00pm-5:30pm - Hall 3 + Hall 2B #58 Website: toponets.github.io