Multitask Learning by Multiwave Optical Diffractive Network

<table class="table-group" id="tab1"><tr><td><table class="table"><tr><td class="thead-hr" colspan="5"><hr/></td></tr><tr class="thead"><td class="align_left">Method</td><td class="align_center">MNIST (%)</td><td class="align_center"><i>λ</i> (THz)</td><td class="align_center">MNIST-Fashion (%)</td><td class="align_center"><i>λ</i> (THz)</td></tr><tr><td class="thead-hr" colspan="5"><hr/></td></tr><tr><td class="align_left">MWDN(PCM)</td><td class="align_center">92.85</td><td class="align_center">0.4</td><td class="align_center">84.33</td><td class="align_center">14.4</td></tr><tr><td class="align_left">DN-MNIST</td><td class="align_center">91.75</td><td class="align_center">0.4</td><td class="align_center">81.13</td><td class="align_center">0.4</td></tr><tr class="table-tr"><td colspan="5"><hr class="tbody-hr"/></td></tr></table></td></tr></table>

<div>The performance of MWDN and DN methods.</div>

Mathematical Problems in Engineering

tab1

Table 1

Table 1: Multitask Learning by Multiwave Optical Diffractive Network 