Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces MetaSeq, a novel sequence-based generative framework for the inverse design of acoustic metamaterials (AMMs) that addresses the challenges posed by acoustic dispersion in broadband target responses. By representing AMMs as structured sequences rather than pixel grids or fixed templates, MetaSeq preserves geometric precision and structural connectivity, enabling effective transformation from target responses to design sequences. The framework's effectiveness is validated through extensive evaluations, demonstrating a 45% reduction in response error compared to the best existing baseline methods.
MetaSeq achieves a 45% reduction in response error for acoustic metamaterial design by transforming the inverse design problem into a structured sequence-to-sequence task.
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.
NUS