公开(公告)号：US20240039244A1

公开(公告)日：2024-02-01

申请号：US17875367

申请日：2022-07-27

Applicant: HEWLETT PACKARD ENTERPRISE DEVELOPMENT LP

Inventor： STANLEY CHEUNG ,  DI LIANG ,  RAYMOND G. BEAUSOLEIL ,  MICHAEL RENNE TY TAN ,  WAYNE VICTOR SORIN

IPC: H01S5/14 ,  H01S5/026 ,  H01S5/10

CPC classification number: H01S5/142 ,  H01S5/0265 ,  H01S5/1071

Abstract: Implementations disclosed herein provide semiconductor resonator based optical multiplexers that achieve enhanced bandwidth range of light emitted therefrom. The present disclosure integrates silicon devices into resonator structures, such as micro-ring resonators, that couples a side mode with a lasing mode and resonantly amplifies coupled light to output light having an enhanced bandwidth with respect to the lasing mode. In some examples, the optical multiplexers disclosed herein include a bus waveguide; a first resonator structure optically coupled to the bus waveguide and comprising an optical amplification mechanism that generates light and a single mode filter to force the generated light into single-mode operation; and a second resonator structure optically coupled to the first resonator structure and comprising a phase-tuning mechanism. The phase-tuning mechanism can be controlled to detune phase of light in the second resonator relative to the light in the first resonator.

公开(公告)号：US20230133722A1

公开(公告)日：2023-05-04

申请号：US17515368

申请日：2021-10-29

Applicant: HEWLETT PACKARD ENTERPRISE DEVELOPMENT LP

Inventor： THOMAS VAN VAERENBERGH ,  PENG SUN ,  MARTIN FOLTIN ,  RAYMOND G. BEAUSOLEIL

IPC: G06N20/20 ,  G06F17/18

Abstract: Systems and methods are provided for creating and sharing knowledge among design houses. In particular, examples of the presently disclosed technology leverage the concepts of meta-optimizing and collaborative learning to reduce the computational burden shouldered by individual design houses using inverse design techniques to find optimal designs in a manner which protects intellectual property sensitive information. Examples may share versions of a central meta-optimizer (i.e. local meta-optimizers) among design houses targeting different (but related) design tasks. A local meta-optimizer can be trained to indirectly optimize a design task by computing hyper-parameters for a design house's private optimization function. The private optimization function may be using inverse design techniques to find an optimal design for a design task. This may correspond to finding a global minimum of a cost function using gradient descent techniques or more advanced global optimization techniques.