NEAT: Neighborhood-Guided, Efficient, Autoregressive Set Transformer for 3D Molecular Generation
By: Daniel Rose , Roxane Axel Jacob , Johannes Kirchmair and more
Autoregressive models are a promising alternative to diffusion-based models for 3D molecular structure generation. However, a key limitation is the assumption of a token order: while text has a natural sequential order, the next token prediction given a molecular graph prefix should be invariant to atom permutations. Previous works sidestepped this mismatch by using canonical orders or focus atoms. We argue that this is unnecessary. We introduce NEAT, a Neighborhood-guided, Efficient, Autoregressive, Set Transformer that treats molecular graphs as sets of atoms and learns the order-agnostic distribution over admissible tokens at the graph boundary with an autoregressive flow model. NEAT approaches state-of-the-art performance in 3D molecular generation with high computational efficiency and atom-level permutation invariance, establishing a practical foundation for scalable molecular design.
Similar Papers
InertialAR: Autoregressive 3D Molecule Generation with Inertial Frames
Machine Learning (CS)
Designs new molecules for medicine and materials.
Learning-Order Autoregressive Models with Application to Molecular Graph Generation
Machine Learning (CS)
Creates new molecules by guessing their parts.
Neighboring Autoregressive Modeling for Efficient Visual Generation
CV and Pattern Recognition
Makes computers create pictures and videos faster.