acceptodds
Desk rejected at ICLR 2027

SE(3)-EQUIVARIANT ACOUSTIC GRAPH OPERATORS: PHYSICAL FOUNDATIONS, STEERABLE MESSAGE PASSING, AND ZERO-SHOT SPATIAL ARRAY GENERALIZATION

Desk rejected at ICLR 2027, so left out of the paper lists and the map.

Abstract

Spatial audio signal processing which include encompassing 3D direction-of-arrival (DOA) estimation, sound source localization, and acoustic scene reconstruction have remained fundamentally constrained by the structural rigidity of deep learning models. Standard multi-channel architectures concatenate microphone array recordings into fixed channel tensors, treating physical sensors as unstructured feature maps and causing catastrophic collapse under spatial rotations, translations, or topological changes in the sensor array. We present the SE(3)-Equivariant Acoustic Graph Network (SE3-AGN), a geometric deep learning framework bridging continuous physical wave mechanics with SE(3)-equivariant neural operators. By representing arbitrary multi-microphone arrays as 3D spatial graphs, SE3-AGN formulates node feature updates using steerable Clebsch–Gordan tensor products over spherical harmonics basis functions, dynamically modulated by cross-correlation phase-lag sequences and inter-sensor Euclidean distances. We introduce an adaptive, task-learned graph sparsification mechanism, physics-informed training constraints, an equivariance-consistency self-supervised pretraining objective, and a symmetry-preserving output head, and we prove mathematically that SE3-AGN guarantees exact equivariance under 3D rigid-body transformations (SE(3)) and invariance under node index permutations (). On synthetic reverberant Room Impulse Response (RIR) simulations, SE3-AGN sets a new state-of-the-art in 3D DOA estimation (1.12 MAE at RT60= 0.2s and 4.61 at RT60= 0.8s), demonstrates zero-shot generalization to unseen microphone array geometries, and exhibits 10 higher sample efficiency over non-equivariant baselines, with real-world evaluation on the LOCATA benchmark confirming robustness across dynamic multi-speaker scenarios.

Primary area:
learning on graphs and other geometries & topologies
Keywords:
SE(3)-equivariant neural networks, geometric deep learning, graph neural networks, spatial audio processing, sound source localization, 3D direction-of-arrival estimation, steerable features, spherical harmonics, Clebsch–Gordan tensor products, zero-shot generalization, microphone array processing, acoustic scene reconstruction

Related papers

Discussion (0)

Sign in to comment.