Scalable Wavelength Arbitration for Microring-based DWDM Transceivers

This paper introduces the concept of autonomous microring arbitration, or \textit{wavelength arbitration}, to address the challenge of multi-microring initialization in microring-based Dense-Wavelength-Division-Multiplexed (DWDM) transceivers. This arbitration is inherently policy-driven, defining critical system characteristics such as the spectral ordering of microrings. Furthermore, to facilitate large-scale deployment, the arbitration algorithms must operate independently of specific wavelength information and be resilient to system variability. Addressing these complexities requires a holistic approach that encompasses the entire system, from device-level variabilities to the transceiver interface - this system-wide perspective is the focus of this paper. To support efficient analysis, we develop a hierarchical framework incorporating an ideal, wavelength-aware arbitration model to examine arbitration failures at both the policy and algorithmic levels. The effectiveness of this approach is demonstrated in two ways: by analyzing the robustness of each policy in relation to device variabilities, and by developing an algorithm that achieves near-perfect alignment with the ideal model, offering superior robustness compared to the traditional sequential tuning method. The simulator code used in this paper is available at \url{https://github.com/wdmsim/wdm-simulator}.

Pretraining Graph Neural Networks for few-shot Analog Circuit Modeling and Design

Being able to predict the performance of circuits without running expensive simulations is a desired capability that can catalyze automated design. In this paper, we present a supervised pretraining approach to learn circuit representations that can be adapted to new circuit topologies or unseen prediction tasks. We hypothesize that if we train a neural network (NN) that can predict the output DC voltages of a wide range of circuit instances it will be forced to learn generalizable knowledge about the role of each circuit element and how they interact with each other. The dataset for this supervised learning objective can be easily collected at scale since the required DC simulation to get ground truth labels is relatively cheap. This representation would then be helpful for few-shot generalization to unseen circuit metrics that require more time consuming simulations for obtaining the ground-truth labels. To cope with the variable topological structure of different circuits we describe each circuit as a graph and use graph neural networks (GNNs) to learn node embeddings. We show that pretraining GNNs on prediction of output node voltages can encourage learning representations that can be adapted to new unseen topologies or prediction of new circuit level properties with up to 10x more sample efficiency compared to a randomly initialized model. We further show that we can improve sample efficiency of prior SoTA model-based optimization methods by 2x (almost as good as using an oracle model) via fintuning pretrained GNNs as the feature extractor of the learned models.