Unverified50% confidenceFactExact time
分子天然具有SE(3)/E(3)等变性，即物理性质不随整体旋转和平移改变；几何深度学习将物理对称先验内嵌于网络结构，使模型参数量更小、泛化更强
1
Sources
50%
Confidence
Long-term
Relevance
7/6/2026
First Seen
Sources
Related Claims
Unverifiede3nn已被应用于分子动力学势能面拟合（如NequIP、MACE等模型）、蛋白质结构预测和材料性质预测65% similarUnverifiedTFN（Tensor Field Networks）、SEGNN和EquiFold等模型将原子视为图节点，用球谐函数编码原子间方向信息，确保预测不依赖输入坐标系60% similarUnverifiedByteDance open-sourced the Protonics biomolecular structure prediction model, which surpasses AlphaFold 3 in core capabilities.60% similarUnverified贝叶斯定理在朴素贝叶斯分类器、贝叶斯神经网络、变分自编码器（VAE）和扩散模型中都有直接应用57% similarUnverifiedAnthropic uses Sparse Autoencoders (SAE) to decompose the polysemantic activations inside the model into purer features55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/114579API
curl https://kongchang.com/api/v1/knowledge/claims/114579MCP
get_claim(id=114579)