Physics-Guided Sequence-Based Generative Framework for Acoustic Metamaterial Inverse Design
2026-06-08 • Sound
SoundArtificial Intelligence
AI summaryⓘ
The authors developed MetaSeq, a new method to design acoustic metamaterials—special materials that control sound—across a wide range of frequencies. Unlike past methods that used fixed shapes or blurry images, MetaSeq represents the material as a detailed sequence showing its exact geometry and how parts connect. This helps designers better match desired sound responses at multiple frequencies without losing precision. The authors trained MetaSeq with a smart combination of learning techniques and showed it works better than existing methods by significantly reducing errors.
Acoustic metamaterialsInverse designBroadband responseAcoustic dispersionSequence-to-sequence modelsReinforcement learningGeometric representationCOMSOLPhysics-based solverGenerative framework
Authors
Yijie Li, Jiahao Xu, Ching-Chih Tsao, Lili Qiu, Jingxian Wang
Abstract
Acoustic metamaterial (AMM) inverse design is particularly challenging for broadband target responses due to acoustic dispersion: a structure that matches the desired response at one frequency may deviate at others, and modifying geometry to improve one sub-band often perturbs neighboring sub-bands. Yet existing broadband inverse-design approaches are either constrained by predefined templates, or rely on image representations that fail to preserve the geometric precision and structural connectivity required by acoustic structures. We present MetaSeq, a physics-guided, sequence-based generative framework for acoustic metamaterial inverse design. At its core, MetaSeq introduces a language that represents each AMM as a structured sequence, rather than as a pixel grid or fixed template. This representation preserves precise geometry, explicitly encodes connectivity, and casts inverse design as a sequence-to-sequence task from target response to structure sequence. MetaSeq further constructs a balanced, high-fidelity dataset with efficient calibration and complexity-based sampling. To address the one-to-many nature of inverse design, MetaSeq combines supervised pretraining with reinforcement learning fine-tuning guided by a physics-based solver and validity checker. Extensive evaluations against COMSOL and five baselines show that MetaSeq reduces response error by 45% over the best baseline.