REVIEW 2 cited by
E(n) Equivariant Normalizing Flows
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
E(n) Equivariant Normalizing Flows
read the original abstract
This paper introduces a generative model equivariant to Euclidean symmetries: E(n) Equivariant Normalizing Flows (E-NFs). To construct E-NFs, we take the discriminative E(n) graph neural networks and integrate them as a differential equation to obtain an invertible equivariant function: a continuous-time normalizing flow. We demonstrate that E-NFs considerably outperform baselines and existing methods from the literature on particle systems such as DW4 and LJ13, and on molecules from QM9 in terms of log-likelihood. To the best of our knowledge, this is the first flow that jointly generates molecule features and positions in 3D.
Forward citations
Cited by 2 Pith papers
-
Drawback of Enforcing Equivariance and its Compensation via the Lens of Expressive Power
Enforcing equivariance reduces expressive power in 2-layer ReLU networks but enlarging the model compensates with proven size bounds and yields lower hypothesis space dimensionality for better generalization.
-
InertialAR: Autoregressive 3D Molecule Generation with Inertial Frames
An autoregressive transformer with inertial-frame tokenization and geometric rotary positional encoding reports state-of-the-art validity and stability on QM9, GEOM-Drugs, and B3LYP, plus strong functional-group-condi...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.