This study presents an innovative approach to implementing vector-matrix multiplication (VMM) using optimized low-index-contrast silicon photonic structures designed via inverse design techniques. Leveraging the inherent parallelism and high bandwidth of silicon photonics, the proposed architecture employs low-index-contrast waveguides to achieve low-loss, high-fidelity signal propagation while maintaining fabrication compatibility with existing CMOS processes. Inverse design algorithms, including adjoint optimization and topology optimization, were utilized to precisely tailor the photonic structures to perform the required linear transformations with minimal crosstalk and insertion loss. Numerical simulations demonstrate that the optimized devices can perform VMM operations with high accuracy, supporting scalability for large-scale photonic computing applications. The results highlight the potential of combining low-index-contrast materials and inverse design to enable efficient, reconfigurable, and compact photonic accelerators for artificial intelligence and signal processing workloads.

Silicon photonics, inverse design, vector-matrix multiplication

Cordaro, A. et al. Solving integral equations in free-space with inverse-designed ultrathin optical metagratings. Nat. Nanotechnol. 18, 365–372 (2023).

Nikkhah, V., Tzarouchis, D. C., Hoorfar, A. & Engheta, N. Inverse-designed metastructures together with reconfigurable couplers to compute forward scattering. ACS Photonics 10, 977–985 (2022).

Silva, A. et al. Performing mathematical operations with metamaterials. Science 343, 160–163 (2014).12

Mohammadi Estakhri, N., Edwards, B. & Engheta, N. Inverse-designed metastructures that solve equations. Science 363, 1333–1338 (2019).

Zhang, H. et al. An optical neural chip for implementing complex-valued neural network. Nat. Commun. 12, 457 (2021).

Camacho, M., Edwards, B. & Engheta, N. A single inverse-designed photonic structure that performs parallel computing. Nat. Commun. 12, 1466 (2021).

Fu, W. et al. Ultracompact meta-imagers for arbitrary all-optical convolution. Light Sci. Appl. 11, 62 (2022).

Pan, D. et al. Laplace metasurfaces for optical analog computing based on quasi-bound states in the continuum. Photon. Res. 9, 1758–1766 (2021).

Zangeneh-Nejad, F., Sounas, D. L., Alù, A. & Fleury, R. Analogue computing with metamaterials.13 Nat. Rev. Mater. 6, 207–225 (2021).

Zhu, T. et al. Plasmonic computing of spatial differentiation. Nat. Commun. 8, 15391 (2017).

Pors, A., Nielsen, M. G. & Bozhevolnyi, S. I. Analog computing using reflective plasmonic metasurfaces. Nano Lett. 15, 791–797 (2015).

Wang, H. et al. Design of compact meta-crystal slab for general optical convolution. ACS Photonics 9, 1358–1365 (2022).

Choutagunta, K., Roberts, I., Miller, D. A. & Kahn, J. M. Adapting Mach–Zehnder mesh equalizers in direct-detection mode-division-multiplexed links. J. Light. Technol. 38, 723–735 (2019).

Knill, E., Laflamme, R. & Milburn, G. J. A scheme for efficient quantum computation with linear optics. Nature 409, 46–52 (2001).

Nikkhah, V., Mencagli, M. J. & Engheta, N. Reconfigurable nonlinear optical element using tunable couplers and inverse-designed structure.14 Nanophotonics 12, 3019–3027 (2023).