{"id":2920,"date":"2026-03-27T15:30:22","date_gmt":"2026-03-27T15:30:22","guid":{"rendered":"https:\/\/cleoconference.org\/?page_id=2920"},"modified":"2026-03-27T15:30:22","modified_gmt":"2026-03-27T15:30:22","slug":"ryan-hamerly","status":"publish","type":"page","link":"https:\/\/cleoconference.org\/ryan-hamerly\/","title":{"rendered":"Ryan Hamerly"},"content":{"rendered":"

Short Course<\/h2><\/div>

CLEO<\/h2><\/div>

The premier international forum for scientific and technical optics.<\/em><\/p>\n<\/div>

REGISTER NOW<\/span><\/a><\/div><\/div><\/div><\/div><\/div>

\u00a0Ryan Hamerly<\/p><\/h3><\/div>

<\/div><\/div><\/div><\/div>

<\/h3><\/div>

Course\u202fTitle: <\/strong>Optical\u00a0<\/span>Machine Learning<\/span><\/span><\/span><\/p>\n

Course\u202fLevel:<\/strong> Advanced Beginner<\/span><\/span>\u00a0<\/span><\/span><\/p>\n

Course\u202fDescription:<\/strong><\/span><\/p>\n

Optics has long been used for communication, but the\u00a0<\/span>recent\u00a0<\/span>rise\u00a0<\/span>of<\/span>\u00a0photonic analog \u201caccelerators\u201d may give it a role in computing as well<\/span>.\u00a0\u00a0<\/span>This course focuses on the principles and recent innovations of optical machine learning systems,\u00a0<\/span>i.e.<\/span>\u00a0optical systems that perform machine learning<\/span>.<\/span>\u00a0\u00a0<\/span>To start, this involves defining the problem, decomposing machine learning into mathematical operations (nonlinearity, matrix multiplication, convolution, etc.) and\u00a0<\/span>identifying<\/span>\u00a0bottlenecks in digital hardware and opportunities for photonic acceleration<\/span>.\u00a0\u00a0<\/span>We highlight the theoretical advantages of optical computing<\/span>,\u00a0<\/span>including high data bandwidth, support for multiplexing (quadrature, mode,\u00a0<\/span>time,\u00a0<\/span>wavelength, polarization),\u00a0<\/span>and low-loss data movement, and define the important figures of merit for any system<\/span>, such as<\/span>\u00a0power consumption, throughput density, latency,\u00a0<\/span>and\u00a0<\/span>bit precision.\u00a0 Based on this theoretical framework, we analyze concrete examples proposed in the literature, including schemes based on programmable interferometers,\u00a0<\/span>microring<\/span>\u00a0weight banks, coherent detection, diffraction, and Fourier optics<\/span>.\u00a0\u00a0<\/span>These systems give a good sense of the\u00a0<\/span>photonic\u00a0<\/span>device<\/span>s\u00a0<\/span>required<\/span>\u00a0for a high-performance machine learning system, which, depending on the scheme, can include phase shifters,\u00a0<\/span>beamsplitters<\/span>, ring resonators, modulators, and photodetectors<\/span>.\u00a0\u00a0<\/span>Material- and device-level design choices are evaluated, which usually align with the choices made in\u00a0<\/span>optical\u00a0<\/span>communications, but sometimes not<\/span>.\u00a0\u00a0<\/span>For perspective, we\u00a0<\/span>also\u00a0<\/span>cover historical approaches to optical\u00a0<\/span>computing<\/span>\u00a0that never worked out (or were supplanted by CMOS<\/span>\u00a0electronics<\/span>) and discuss why recent technological innovations\u00a0<\/span>in photonics have leveled the playing field and may make optical machine learning\u00a0<\/span>viable<\/span>\u00a0in the future.<\/span><\/span>\u00a0<\/span><\/p>\n

Benefits\u202fand\u202fLearning\u202fObjectives:<\/strong>\u00a0<\/span><\/p>\n

    \n
  • Define and quantify the benefits of using optics for computing<\/span>\u00a0<\/span><\/li>\n<\/ul>\n
      \n
    • Compare the advantages and drawbacks of free-space and on-chip architectures<\/span>\u00a0<\/span><\/li>\n<\/ul>\n
        \n
      • List and define the figures of merit for an optical machine learning system<\/span>\u00a0<\/span><\/li>\n<\/ul>\n
          \n
        • Describe factors that affect or limit the scalability of optical computing hardware<\/span>\u00a0<\/span><\/li>\n<\/ul>\n
            \n
          • List ways that computation hardware can\u00a0leverage\u00a0optical degrees of freedom,\u00a0e.g.\u00a0amplitude, phase, mode, wavelength, polarization<\/span>\u00a0<\/span><\/li>\n<\/ul>\n
              \n
            • Distinguish optical machine learning from analog electronic approaches<\/span>\u00a0<\/span><\/li>\n<\/ul>\n
                \n
              • Calculate\u00a0the energy consumption of an optical computing system<\/span>\u00a0<\/span><\/li>\n<\/ul>\n

                Intended\u202fAudience:<\/strong><\/p>\n

                Open to students, postdocs, and professionals with a background in optics \/ photonics, who are interested in researching or developing optical machine-learning systems or<\/span>\u00a0simply keeping up with the field<\/span>.\u00a0\u00a0<\/span>Knowledge of machine learning,<\/span>\u00a0computer systems, and<\/span>\u00a0<\/span>integrated photonics is a plus but not\u00a0<\/span>required<\/span>.<\/span><\/span>\u00a0<\/span><\/p>\n

                Instructor\u202fBiography:\u00a0<\/strong><\/p>\n

                My background is in theoretical and quantum physics, but recently I have focused on integrated photonic systems for computing.\u00a0\u00a0I co-lead the optical AI hardware teams at the Quantum Photonics Group at MIT (Prof. Englund) and NTT Research.\u00a0\u00a0In the past, I worked on OPO coherent\u00a0Ising\u00a0machines, coherent feedback control, and gravitational wave physics.\u00a0\u00a0B.S. Caltech 2010, PhD Stanford 2016.<\/span>\u00a0<\/span><\/p>\n

                \u00a0<\/span><\/p>\n<\/div><\/div><\/div>

                <\/h4><\/div>

                <\/h4><\/div><\/div><\/div><\/div><\/div><\/p>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":2,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"100-width.php","meta":{"footnotes":""},"class_list":["post-2920","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/pages\/2920","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/comments?post=2920"}],"version-history":[{"count":2,"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/pages\/2920\/revisions"}],"predecessor-version":[{"id":2922,"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/pages\/2920\/revisions\/2922"}],"wp:attachment":[{"href":"https:\/\/cleoconference.org\/wp-json\/wp\/v2\/media?parent=2920"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}