AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

HelioX: A GPU-Native Framework for Simulation and Training of Biophysically Detailed Networks

JUNFENG LU, Zijie Yu, Shaoyang Cui, Gan He, Ruiqin Xiong, Kai Du, Tiejun Huang

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Brain-inspired network comparison

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use HelioX instead of CoreNEURONsimulator · Single L5PCsimulation time: 2.99 -> 1.33Improved
use HelioX instead of CoreNEURONsimulator · Active 500 HPCsimulation time: 8.04 -> 3.92Improved
use HelioX instead of JAXLEYsimulator · Passive 500 HPCsimulation time: 10.30 -> 3.30Improved
add 5x learning-time downsampling for gradient computationHelioX · BAAIWorm circuit fittingtime per epoch (s): 5100 -> 60Improved
add 5x learning-time downsampling (ablation)HelioX · BAAIWorm circuit fittingpeak GPU memory (GB): 5.2 -> 1.6Improved
use HelioX instead of CoreNEURON for trainingtraining framework · BAAIWorm circuit fittingpeak GPU memory (GB): 47.4 -> 5.2Improved
use HelioX instead of JAXLEY for MLP trainingtraining framework · MNIST biophysical MLP (3-layer)training time (s): 60.41 -> 5.06Improved
use HelioX instead of DeepDendrite for MLP trainingtraining framework · MNIST biophysical MLP (3-layer)training time (s): 9.47 -> 5.06Improved
use HelioX instead of CoreNEURON for MLP trainingtraining framework · MNIST biophysical MLP (3-layer)training time (s): 220.99 -> 5.06Improved
use HelioX instead of CoreNEURONtraining framework · MNIST biophysical MLP (3-layer)peak GPU memory (MiB): 1079 -> 1121No clear effect
use HelioX instead of JAXLEYtraining framework · MNIST biophysical MLP (3-layer)peak GPU memory (MiB): 2219 -> 1121Improved
use HelioX instead of DeepDendritetraining framework · MNIST biophysical MLP (3-layer)peak GPU memory (MiB): 1321 -> 1121Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy