NEAT transformer generates 3D molecules with state-of-the-art speed and accuracy

Researchers have developed NEAT, a novel autoregressive set transformer designed for 3D molecular generation. Unlike previous methods that rely on sequential atom ordering, NEAT treats molecules as sets and uses a neighborhood-guided training strategy to ensure permutation invariance. This approach allows the model to learn an order-agnostic distribution over tokens, leading to state-of-the-art generation quality on datasets like QM9 and GEOM-Drugs while also being significantly faster than existing methods. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a novel permutation-invariant approach for 3D molecular generation, potentially accelerating drug discovery and materials science research.

RANK_REASON This is a research paper detailing a new model architecture for molecular generation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• NEAT
• Transformer
• QM9
• GEOM-Drugs
• arXiv
• Daniel Rose

• paper
• model release