Research Project: Integrated Photonic Neural Networks with Arbitrary Capabilities
Loading...
Contributors
Funders
ID
EC.00151
Authors
Mağden, Emir Salih
Faculty Member
Publications
Deep photonic network platform enabling arbitrary and broadband optical functionality
(Nature Portfolio, 2024) Amiri, Ali Najjar; Görgülü, Kazım; Mağden, Emir Salih; Vit, Aycan Deniz; ; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; Yes; College of Engineering; GRADUATE SCHOOL OF SCIENCES AND ENGINEERING
Expanding applications in optical communications, computing, and sensing continue to drive the need for high-performance integrated photonic components. Designing these on-chip systems with arbitrary functionality requires beyond what is possible with physical intuition, for which machine learning-based methods have recently become popular. However, computational demands for physically accurate device simulations present critical challenges, significantly limiting scalability and design flexibility of these methods. Here, we present a highly-scalable, physics-informed design platform for on-chip optical systems with arbitrary functionality, based on deep photonic networks of custom-designed Mach-Zehnder interferometers. Leveraging this platform, we demonstrate ultra-broadband power splitters and a spectral duplexer, each designed within two minutes. The devices exhibit state-of-the-art experimental performance with insertion losses below 0.66 dB, and 1-dB bandwidths exceeding 120 nm. This platform provides a tractable path towards systematic, large-scale photonic system design, enabling custom power, phase, and dispersion profiles for high-throughput communications, quantum information processing, and medical/biological sensing applications. An efficient and physically accurate platform is required to rapidly design high-performance integrated photonic devices. Here, the authors present a scalable framework for creating on-chip optical systems with complex and arbitrary functionality.
Layout-aware and fabrication-tolerant deep photonic networks
(Institute of Electrical and Electronics Engineers Inc., 2025) Vit, Aycan Deniz; Görgülü, Kazım; Danış, Bahrem Serhat; Rzayev, Ujal; Daşdemir, Ahmet Onur; Mağden, Emir Salih; Görgülü, Kazım; Amiri, Ali Najjar ; Maǧden, Abdullah ; Graduate School of Sciences and Engineering; Department of Electrical and Electronics Engineering; KUIS AI (Koç University & İş Bank Artificial Intelligence Center); Yes; Daşdemir, Ahmet Onur; Danış, Bahrem Serhat; Rzayev, Ujal; Vit, Aycan Deniz; GRADUATE SCHOOL OF SCIENCES AND ENGINEERING; College of Engineering; Research Center
We propose a systematic approach to design complex and universally capable deep photonic networks with inherent robustness to fabrication-induced variations. We first create a framework that incorporates layers of variation-aware, custom-designed Mach-Zehnder interferometers and virtual wafer maps to optimize integrated photonic devices under fabrication imperfections. Using this framework, we propose designs for silicon-based, broadband 50/50 splitters and 5% power taps demonstrating transmissions within ±2% and ±1.5% of their respective targets across the 1.5-1.6 μm spectrum, even in the presence of fabrication variations of ±15 nm in waveguide width and ±10 nm in thickness. Due to their consistent performance under fabrication-induced variations, we show that the potential fabrication yield for these tolerant devices is drastically higher than traditional devices. Our results underscore the effectiveness of the deep photonic network architecture in building layout-aware photonic systems, showcasing a realistic pathway towards creating inherently robust geometries with consistent performance for maximum yield and device reliability in future photonic applications.
